Forthcoming and Online First Articles

International Journal of Reasoning-based Intelligent Systems

International Journal of Reasoning-based Intelligent Systems (IJRIS)

Forthcoming articles have been peer-reviewed and accepted for publication but are pending final changes, are not yet published and may not appear here in their final order of publication until they are assigned to issues. Therefore, the content conforms to our standards but the presentation (e.g. typesetting and proof-reading) is not necessarily up to the Inderscience standard. Additionally, titles, authors, abstracts and keywords may change before publication. Articles will not be published until the final proofs are validated by their authors.

Forthcoming articles must be purchased for the purposes of research, teaching and private study only. These articles can be cited using the expression "in press". For example: Smith, J. (in press). Article Title. Journal Title.

Articles marked with this shopping trolley icon are available for purchase - click on the icon to send an email request to purchase.

Online First articles are published online here, before they appear in a journal issue. Online First articles are fully citeable, complete with a DOI. They can be cited, read, and downloaded. Online First articles are published as Open Access (OA) articles to make the latest research available as early as possible.

Open AccessArticles marked with this Open Access icon are Online First articles. They are freely available and openly accessible to all without any restriction except the ones stated in their respective CC licenses.

Register for our alerting service, which notifies you by email when new issues are published online.

We also offer which provide timely updates of tables of contents, newly published articles and calls for papers.

International Journal of Reasoning-based Intelligent Systems (43 papers in press)

Regular Issues

  • Identification and Detection of Brain Tumour Using Deep Learning-based Classification MRI Applied Using Neural Network and Machine Learning Algorithm   Order a copy of this article
    by Biswaranjan Mishra, Kakita Gopal, Srikant Patnaik, Bijay Paikaray 
    Abstract: Human brain is considered as the most sophisticated part of the body and used to consist of several neurons and biological components. The normal brain is usually functioning at an aspect of 93% for a healthy human. A brain tumour is a common disease nowadays and this disease usually leads to the accumulation of aberrant cells in certain brain tissues. Which may cause the formation of dump cells in the brain. One of the most valuable approaches is the MRI images which can identify the various stages for the detection of the tumour. Here a variety of feature extraction and classification techniques are available and MRI pictures are used to identify brain tumours. Here in this paper, the convolutional neural network approach is discussed where the high-accuracy image classification technique for early tumour detection is used.
    Keywords: MRI; human brain; tumour; CNN; and Glioma; RNN.
    DOI: 10.1504/IJRIS.2023.10051697
     
  • A Recommendation method for online learning resources of mathematics courses based on feature graph clustering   Order a copy of this article
    by Zhixia Duan, Na Zhao 
    Abstract: Due to the low comprehensiveness of the traditional method to the analysis of learners’ needs, the degree of fitting between the recommended learning resources and the actual needs of learners is low. To solve this problem, a recommendation method for online learning resources of mathematics courses based on feature graph clustering is proposed. Construct learner feature map from cognitive level and learning preference, and analyse their attribute characteristics, behaviour characteristics and learning characteristics. Then on the basis of clustering processing, resources with the same clustering characteristics are matched as the recommendation target. The test results show that the satisfaction of the recommendation results of this method is always stable at more than 90.0%, the maximum and minimum F1-score values are 0.52 and 0.46, respectively, and it has high stability, which is obviously better than the traditional method.
    Keywords: online learning resources; cognitive level; learning preferences; characteristic atlas; feature similarity; clustering processing; fitting degree.
    DOI: 10.1504/IJRIS.2023.10051721
     
  • A fast identification method of UAV inspection image target for substation equipment   Order a copy of this article
    by Wen Kang, You Li, DeKai Liu, JieXue Zheng, Yisheng Yu 
    Abstract: Aiming at the problems of low accuracy, long time and poor recognition effect of traditional methods, a fast identification method of UAV inspection image target for substation equipment is proposed. Firstly, Gabor filter is used to extract the target features of UAV patrol image of substation equipment, and the UAV patrol image is segmented. Then, the binary image is processed, and the constraint criterion of image reconstruction is established to realise image reconstruction. Finally, all the image blocks are normalised, and the target model is constructed on the basis of prior knowledge to match and find the corresponding target, so as to realise the rapid recognition of the image target of UAV patrol inspection of substation equipment. The experimental results show that the target recognition accuracy of this method is always higher than 70%, and the minimum recognition time is only 3.26s. The target recognition effect is good.
    Keywords: substation equipment; UAV patrol inspection; image recognition; Gabor filter; image reconstruction; prior knowledge; image segmentation; binarisation.
    DOI: 10.1504/IJRIS.2023.10051722
     
  • The fast multi threshold color image segmentation based on decision rough set   Order a copy of this article
    by Yanting Cao 
    Abstract: In order to improve the speed and quality of multi-threshold colour image segmentation, a new multi-threshold colour image segmentation method based on decision rough set is proposed. Firstly, the RGB colour space is constructed, and the average brightness value is used to calculate the position segmentation points of the global histogram of the multi-threshold image. Then, the output membership value is obtained by decision rough set, and the multi-threshold image colour channel compensation and attenuation information are determined. Finally, the Gauss function is used to determine the segmentation position of each channel of the multi threshold image, and the fast multi-threshold colour image segmentation is completed. The experimental results show that the segmentation coefficient of this method is 0.868, the maximum PSNR value is 0.982 and the minimum MSE value is 0.003.
    Keywords: multi-threshold colour image; RGB colour space; histogram equalisation; decision rough set; image segmentation.
    DOI: 10.1504/IJRIS.2023.10051723
     
  • The control of lower limb rehabilitation robot with multi-pose based on man-machine coupling and fuzzy PID   Order a copy of this article
    by Xiaomei Kang 
    Abstract: In order to accurately control the joint displacement and joint angle of the lower limb rehabilitation robot, a control method for lower limb rehabilitation robot with multi-pose based on man-machine coupling and fuzzy PID is proposed. According to the theory of lower limb rehabilitation medicine, the human lower limb model is constructed. The FSR membrane pressure sensor is selected to obtain the man-machine contact force. Based on the man-machine coupling theory, the interaction requirements between the patient and the lower limb rehabilitation robot are analysed. The fuzzy PID controller is designed to control the lower limb rehabilitation robot with multi-pose on the basis of fuzzy rules to help the patient complete the rehabilitation training. The experimental results show that the proposed method can accurately control the joint displacement and joint angle of the lower limb rehabilitation robot, and has good positioning accuracy.
    Keywords: man-machine coupling; fuzzy PID controller; lower limb rehabilitation robot with multi-posture; kinematic analysis; FSR membrane pressure sensor.
    DOI: 10.1504/IJRIS.2023.10051724
     
  • The Individualized Travel Route Selection Based on Dynamic Transfer Graph   Order a copy of this article
    by Xiao-Qin Geng, Yun-Duo Wang 
    Abstract: In order to solve the problems of low selection accuracy, high cost and low tourist satisfaction in traditional methods, this paper proposes an individualised travel route selection method based on dynamic transfer graph. The granular time set is introduced to map the binary set of any event in the scenic spot to the triplet, so as to realise the construction of the dynamic transfer graph of the scenic spot. TF-IDF technology is used to calculate the fitting degree between tourists’ demand and scenic spots, so as to preliminarily select scenic spots’ tourism routes, complete the secondary screening of scenic spots’ tourism routes according to the dynamic transfer map, and realise individualised travel route selection. The test results show that the maximum selection accuracy of the proposed method is 98.3 and the minimum is 95.6%, the cost is low, and the user satisfaction is the highest, reaching 9.7.
    Keywords: dynamic transfer graph; personalisation; travel route; granular time set; dynamic transfer cycle; TF-IDF technology.
    DOI: 10.1504/IJRIS.2023.10051778
     
  • A balanced allocation method of learning resources in smart classroom based on regional clustering   Order a copy of this article
    by Guangquan Zhou, Yaqi He, Pengwei Li 
    Abstract: In view of the problems of balanced allocation and large error in traditional methods, this paper designs a balanced allocation method of learning resources in smart classroom based on regional clustering. First, the resource data is mapped to the cloud computing network, and the weighted undirected graph is used to complete the data collection. Then, the distribution characteristics of resources are extracted according to the continuity of resource data. Finally, using the regional clustering algorithm, after determining the data core points, according to the distance between the various resource data and the core points, the objective function is used to construct the balanced allocation algorithm. Experimental results show that the allocation balance coefficient of this method is always kept at about 0.9 and the allocation error is kept at about 1%, which indicates that this method can improve the effect of balanced allocation of learning resources.
    Keywords: regional clustering; smart classroom; allocation of learning resources; naive Bayesian algorithm; decision chart; local density.
    DOI: 10.1504/IJRIS.2023.10051981
     
  • Speech unlocking information recognition method of intelligent electronic lock based on acoustic spectrum characteristics   Order a copy of this article
    by Zhifeng Lu, Lingling Ma 
    Abstract: In order to overcome the problems of low recognition accuracy and long recognition time in traditional speech unlocking information recognition methods, a speech unlocking information recognition method of intelligent electronic lock based on acoustic spectrum characteristics is proposed. Firstly, the voice unlocking information of intelligent electronic lock is collected by digital filter, and the unlocking information is enhanced by spectral subtraction. Secondly, the acoustic spectrum of unlocking information after enhanced processing is calculated, and the characteristics of voice unlocking information of intelligent electronic lock are extracted based on the calculated acoustic spectrum. Finally, the short-time autocorrelation function is used to detect the unlocking information endpoint and recognise the voice unlocking information of intelligent electronic lock. The simulation results show that the recognition time of this method is short, the missing recognition rate is low, and the recognition accuracy is high, and the shortest recognition time is only 2.5 s.
    Keywords: acoustic spectrum characteristics; intelligent electronic lock; voice unlocking; spectral subtraction; short-time autocorrelation function.
    DOI: 10.1504/IJRIS.2023.10052189
     
  • Vulnerability assessment method of network topology structure based on maximum likelihood function   Order a copy of this article
    by Qing Li 
    Abstract: In order to improve the accuracy of network vulnerability assessment results, reduce the node loss rate, a network topology vulnerability assessment method based on maximum likelihood function is proposed. The network characteristics and network topology are analysed. The reliability of the network topology is estimated by using the maximum likelihood function according to the analysis results. An assessment index set is constructed according to the reliability assessment results, and each assessment index in the assessment index set is standardised. At the same time, a fuzzy assessment matrix and a judgement scale matrix of network topology vulnerability are constructed, and the weight of each assessment index is calculated. Establish a network topology vulnerability assessment model to achieve vulnerability assessment. The experimental results show that the minimum assessment time of the proposed method is 1.93 s, the node loss rate is always lower than 2%, and the assessment accuracy is high.
    Keywords: maximum likelihood function; network topology; vulnerability assessment; constraints; normalisation.
    DOI: 10.1504/IJRIS.2023.10052190
     
  • An Efficient data transmission method of IOT terminal based on cloud and fog hybrid computing   Order a copy of this article
    by Benjie Wei  
    Abstract: In order to overcome the problems of long transmission delay and small throughput in the terminal data transmission method, this paper proposes an efficient terminal data transmission method of the internet of things based on cloud and fog hybrid computing. Firstly, the architecture of IOT terminal is analysed to determine the random selection coefficient of coding. Secondly, the coding information is solved by Gauss elimination method. Then, CSI modulation is used to remove the terminal interference data. Finally, the channel allocation probability is determined, and the transmission channel allocation is realised through the cloud and fog hybrid calculation method, so as to realise the efficient transmission of the terminal data of the internet of things. The experimental results show that the transmission delay of the design method is always controlled below 3 s, and the network throughput is always higher than 50 MB s
    Keywords: cloud and fog mixing calculation; IOT terminal data; efficient transmission; code; CSI modulation.
    DOI: 10.1504/IJRIS.2023.10052191
     
  • A narrative review on the characterization of automated human emotion detection systems using biomedical sensors and machine intelligence   Order a copy of this article
    by Stobak Dutta, Brojo Mishra, Anirban Mitra, Amartya Chakraborty 
    Abstract: In our day-to-day life, emotion plays an essential role in decision-making and human interaction. For many years, psychologists have been trying to develop many emotional models to explain the human emotional or affective states. Automated emotion recognition is a popular research problem increasingly utilised in marketing, education, health sector, and human-robot interaction. There are different ways of emotion recognition, namely uni-modal and multi-modal solutions. However, it depends upon the purpose for which it is to be used. The primary focus of this paper is to provide a detailed review of the existing literature in this domain with the help of three verticals. Initially, the standard databases used for elicitation of human emotions are discussed in brief. Next, the different sensing approaches used to gather physiological signals are discussed. Finally, a thorough review of the state-of-the-art is given, with reference to the emotional states from the circumplex model of valence-arousal plane.
    Keywords: human emotions; emotion perception; electroencephalography; EEG; galvanic skin response; GSR; electromyographic signal; EMG; skin temperature measurements; SKT.
    DOI: 10.1504/IJRIS.2023.10052192
     
  • Machine Learning Approach to Roof Fall Risks Classification in UG Mines using Adaboost and XGboost Incorporating Transfer Learning Technique   Order a copy of this article
    by Jitendra Pramanik, Bijay Paikaray, Singam Jayanthu, Abhaya Kumar Samal 
    Abstract: Roof stability in underground coal mines is critical in commanding mine productivity as well as miners’ safety. From this perspective, it is a distinctive challenge to provide a safe working environment along with uncompromised productivity and uninterrupted mining operations. Tested over time, machine learning techniques have evolved as a trusted tool in delivering successful outcomes and in providing trustworthy solutions to many real-life problems in various domains of application that can be safely extended to be adopted in this context. The prime objective of this paper is to propose a transfer learning technique-based approach to classify the occurrence of sudden roof fall based on the available roof sag data. The potency of AdaBoost classification algorithms like decision tree, Gaussian Na
    Keywords: transfer learning techniques; roof fall classification; machine learning techniques; AdaBoost classification; XGBoost.
    DOI: 10.1504/IJRIS.2023.10052334
     
  • Workflow scheduling method under unbalanced conditions based on dynamic allocation algorithm   Order a copy of this article
    by Wenyan Zhao, Lixia Hou, Jie Zhao 
    Abstract: In order to improve the efficiency and fairness of workflow scheduling, consider the data correlation between task nodes, and adapt to different scenarios, an unbalanced workflow scheduling method based on dynamic allocation algorithm is proposed. The workflow structure is transformed to calculate the earliest and the latest start time of the node task. The workflow stage is divided, and the input parameters of the task are dynamically adjusted according to the output data amount of each stage of the workflow. The workflow scheduling model is established. Taking the total length of workflow scheduling as the solution goal, the global optimal solution of the whole workflow scheduling process is calculated and obtained. The experimental results show that the unfairness coefficient of this method is less than 0.15%, and the time span can be controlled within 150 s, which shows that this method is more effective and can ensure the running ability of workflow.
    Keywords: dynamic allocation algorithm; workflow stage; non-equilibrium condition; workflow scheduling.
    DOI: 10.1504/IJRIS.2023.10052499
     
  • Study on Machine English Translation Error Identification Based on Naive Bayesian Algorithm   Order a copy of this article
    by Zhongping Yao 
    Abstract: In order to improve the effect and accuracy of machine English translation error identification and shorten the time of machine English translation error identification, a machine English translation error identification method based on naive Bayesian algorithm is proposed. Naive Bayesian algorithm is used to extract the information features of machine English translation. Combined with the automatic translation evaluation method, the results of machine English translation are judged. According to the judgment results, the maximum entropy classifier is selected to identify the error types of machine English translation information, so as to realise the error identification of machine English translation. The experimental results show that the proposed method has a good effect on machine English translation error identification, can effectively improve the accuracy of machine English translation error identification and shorten the time of machine English translation error identification.
    Keywords: naive Bayesian algorithm; maximum entropy classifier; automatic translation evaluation; machine English translation; translation error identification.
    DOI: 10.1504/IJRIS.2023.10052500
     
  • Classification method of popular music score style based on SVM   Order a copy of this article
    by Qiang Tuo, Xiaoming Zhao 
    Abstract: In order to improve the classification accuracy of popular music score style, this paper proposes a new classification method of popular music score style based on SVM. Firstly, the fractional spectral subtraction algorithm is used to enhance the popular music score signal. Secondly, according to the unique characteristics of the striking component and the harmony part in the pop music spectrum, the striking part and the harmony part in the pop music spectrum are separated by means of energy peak removal and peak frequency filtering. Finally, taking the striking part and harmony part of the separated pop music spectrum as input, the SVM algorithm is used to realize the style classification of pop music. The experimental results show that this method can enhance the popular music score signal, and can effectively classify different styles of popular music score, with a classification accuracy of 99.4%.
    Keywords: Support vector machine; Pop music; Music score style; Classification method; Kernel function; Classification function.
    DOI: 10.1504/IJRIS.2023.10052718
     
  • A New Double Attention Decoding Model Based on Cascade RCNN and Word Embedding Fusion for Chinese-English Multimodal Translation   Order a copy of this article
    by Haiying Liu 
    Abstract: Traditional multimodal machine translation (MMT) is to optimise the translation process from the source language to the target language with the help of important feature information in images. However, the information in the image does not necessarily appear in the text, which will interfere with the translation. Compared with the reference translation, mistranslation can be appeared in the translation results. In order to solve above problems, we propose a double attention decoding method based on cascade RCNN to optimise existing multimodal neural machine translation models. The cascade RCNN is applied to source language and source image respectively. Word embedding is used to fuse the initialisation and the semantic information of the dual encoder. In attention computation process, it can reduce the focus on the repetitive information in the past. Finally, experiments are carried out on Chinese-English test sets to verify the effectiveness of the proposed method. Compared with other state-of-the-art methods, the proposed method can obtain better translation results.
    Keywords: multimodal machine translation; MMT; double attention decoding; cascade RCNN; word embedding fusion.
    DOI: 10.1504/IJRIS.2023.10052939
     
  • Machine Learning and Artificial Neural Network for Data Mining Classification and Prediction of Brain Diseases   Order a copy of this article
    by Afrah Salman Dawood 
    Abstract: Recently artificial intelligence (AI), machine learning (ML) and deep learning (DL) got the most of researchers’ attention in different aspects of computing applications and areas such as classification, prediction, etc. However, the development of data mining and its availability assists in the performance evaluation of such models. In this research, different intelligent algorithms (XGBoost, decision tree (DT), random forest, K-NN, ANN, LDA and AdaBoost) were implemented and tested for evaluation and performance. It is worth mentioning that ANN is a DL algorithm while all other algorithms lie in the field of ML. These models were implemented on a combination of Kaggle stroke and Parkinson brain diseases dataset. The performance evaluation of these algorithms computed according to different metrics including precision, recall, f1-score, AUC and accuracy. The accuracy of these models was 97.04% for XGBoost, 95.2% for DT, 97.06% for random forest, 95.02% for K-NN, 95.03% for SVM, 94.95% for logistic regression, 93% for ANN, 94.23% for LDA and 94.71% for AdaBoost. The highest AUC performance was 93.35% for logistic regression. Finally, a comparison in performance with other research was evaluated in terms of accuracy.
    Keywords: data mining; data analysis; brain diseases; artificial intelligence; AI; machine learning; ML; deep learning; DL; big data ANN.
    DOI: 10.1504/IJRIS.2023.10052940
     
  • A Block-Based Fragile Watermarking Scheme for Digital Image Authentication and Tamper Recovery   Order a copy of this article
    by Monsalisa Swain, Debabala Swain 
    Abstract: Unauthorised access and modification of multimedia content are on the rise with the huge increase in digital communication and multimedia information exchange over the web. For the transmitted images to remain protected and authentic, fraud identification and restoration procedures are required. In light of the aforementioned difficulties, a unique self-embedding block-based fragile watermarking method is introduced with enhanced tamper identification and restoration abilities. In this presented method of watermarking, the cover image is split into non-overlapping 4 ? 4 block segments. Seven MSBs of each pixel in the block are used to create the watermark data. The mapping number for each block, where the recovery information is embedded, is created using a key value for that block. Various tampering rates and the number of altered images are used to evaluate the proposed methodology. The SSIM and PSNR values of the recovered image illustrate the uniqueness and efficacy of the method.
    Keywords: spatial domain; fragile watermarking; least significant bit; LSB; singular value decomposition; authentication; image recovery.
    DOI: 10.1504/IJRIS.2023.10052941
     
  • Study on Color distortion correction method of two-dimensional art image based on Naive Bayes   Order a copy of this article
    by Quan Gan, Yucai Zhou 
    Abstract: Aiming at the problems of low correction accuracy and long correction time in the traditional two-dimensional art image colour distortion correction method, a two-dimensional art image colour distortion correction method based on naive Bayes is proposed. Through the RGB image sensor, the brightness pixels are relatively evenly distributed on the histogram through the conversion of four links: world coordinate system, camera coordinate system, photosensitive device coordinate system and pixel coordinate system. Binarisation is carried out based on naive Bayes. According to the binarisation results, the salient regions of two-dimensional art images are extracted. Based on the obtained key regions, the multi-scale retinal algorithm with colour restoration is used to correct the colour distortion of two-dimensional art images. The experimental results show that the proposed method has the highest accuracy and the shortest correction time.
    Keywords: naive Bayes; 2D art image; colour distortion; histogram equalisation; multi-scale retinal algorithm.
    DOI: 10.1504/IJRIS.2023.10053037
     
  • An accurate recommendation method of tourism route based on crawler technology   Order a copy of this article
    by Hui Wang 
    Abstract: A new accurate tourism route recommendation method based on crawler technology is proposed in order to solve the problems of low coverage and long time-consuming of traditional tourism route recommendation methods. Firstly, the process of crawler technology was analysed, and the crawler technology was used to crawl and collect the publicly available tourism demand data in the tourism website. Secondly, the correlation matrix between tourists and tourist routes was constructed, and the interest characteristics of tourist routes were calculated according to the feature sequence. Finally, the candidate tourist route set is generated according to the interest correlation parameters between tourists’ travel time and scenic spots, sorted according to the interest, and the route most in line with the interest was recommended to tourists. The experimental results show that the proposed method is compared with the traditional recommendation methods. Its recommended route has a higher coverage and shorter recommendation time.
    Keywords: reptile technology; tourist routes; accurate recommendation; degree of interest.
    DOI: 10.1504/IJRIS.2023.10053457
     
  • Software running defect recognition method based on brainstorming optimization algorithm   Order a copy of this article
    by Xiaojuan Guo, Fumin Shang, Jiali Chen 
    Abstract: In order to improve the accuracy of software running defect recognition, this paper proposes a new method of software running defect recognition based on brainstorming optimization algorithm. Firstly, K nearest neighbor algorithm is used to cluster and sample software operation data, and Tomek chain removal algorithm is used to remove duplicate data from software operation data. Secondly, after extracting the optimal values of scale factor, displacement factor and inter layer connection weight with brainstorming optimization algorithm, wavelet neural network is trained to complete the optimization of wavelet neural network. Finally, the software operation data after de duplication is input into the optimized wavelet neural network, and the output result is the defect identification result. The experimental results show that the number of software program defects identified by this method is 0 in a specified time, and it has the ability to accurately identify software running defects.
    Keywords: Brainstorming optimization algorithm; Software operation; Defect recognition; Wavelet neural network.
    DOI: 10.1504/IJRIS.2023.10053458
     
  • The computer network security situation awareness based on decision tree algorithm   Order a copy of this article
    by Dehua Kong, Lin Lu, Nian Xiao 
    Abstract: Aiming at the problems of security awareness error and high channel collision rate in computer network security awareness, a computer network security situation awareness method based on decision tree algorithm is proposed. First, analyse the computer network architecture and security requirements; Then, the security situation index classification model is constructed, the network status data is collected, and the data features are extracted using the artificial fish swarm algorithm; Finally, the maximum information gain attribute is used as a sub node of the decision tree to generate a decision tree. Combined with cosine similarity, computer network security situational awareness is realised. The experimental results show that the correlation of the impact indicators of this method is higher than 0.5, the absolute percentage error of security situation awareness is lower than 0.5%, and the channel collision rate is 0~0.10
    Keywords: decision tree algorithm; computer network; security situation awareness.
    DOI: 10.1504/IJRIS.2023.10053459
     
  • Table Tennis Player Expression Recognition Method Based on Gabor Multi directional Feature Fusion   Order a copy of this article
    by Xingbo Zhou, Junmin Wang 
    Abstract: In order to improve the recognition rate of table tennis players’ expressions and reduce the recognition difference, this paper proposes an expression recognition method based on Gabor multi-directional feature fusion. After extracting multi-scale geometric features of facial expression image, the parameters of Gabor filter are optimised, multi-scale feature fusion and filtering are performed on the image, and block histogram features are extracted. The optimised multi-scale features are input into the generated confrontation network model to realise the recognition of table tennis players’ expressions. Experimental results show that the maximum recognition rate of the method can reach 98.7%, and the minimum recognition difference is only 0.871. The feature results of the image in five scales and eight directions can be obtained, which shows that the method can accurately output the facial expression recognition results.
    Keywords: Gabor; multi-scale feature fusion; table tennis players; expression recognition.
    DOI: 10.1504/IJRIS.2023.10053460
     
  • A privacy protection method for IoT nodes based on convolutional neural network   Order a copy of this article
    by Yuexia Han, Di Sun, Yanjing Li 
    Abstract: In order to improve the security of internet of things, a privacy protection method of internet of things nodes based on convolutional neural network is proposed. Firstly, construct the flow model of IoT network nodes, and use the ant colony algorithm to solve the model to obtain the current flow data of IoT nodes. Secondly, a convolutional neural network model is established to identify abnormal nodes in the internet of things. Finally, the privacy protection strategy of k-anonymous IoT nodes based on the average degree of nodes is adopted to protect the privacy of IoT abnormal nodes. The experimental results show that the method can effectively extract the node traffic before and after the attack on the internet of things, and the deviation value is only 2 kb/s; the identification results are more accurate, and the privacy of the internet of things nodes can be effectively protected.
    Keywords: convolutional neural network; internet of things; IoT; node privacy; protection method; anonymity.
    DOI: 10.1504/IJRIS.2023.10053461
     
  • Personalized recommendation method of tourist routes based on multi-constraint and multi-objective   Order a copy of this article
    by Qiang Yang, Yueyun Lai 
    Abstract: In order to improve the accuracy of personalised recommendation of tourist routes and shorten the time-consuming of personalised recommendation, a multi-constraint and multi-objective based personalised recommendation method for tourist routes is proposed. Firstly, collect tourist attraction information, calculate the arrival time, stay time, travel cost, valid itinerary and other parameters between tourist attractions, and calculate the scenic spot score. Secondly, under the constraints of multiple objective parameters of travel time, itinerary and cost, the objective function of personalised recommendation of travel routes is constructed. Finally, the greedy algorithm is used to solve the objective function of personalised recommendation of tourist routes, and the personalised recommendation results of tourist routes are obtained. The experimental results show that the method in this paper can improve the accuracy of personalised recommendation on the basis of shortening the time-consuming of personalised recommendation, and the recommendation accuracy always remains above 90%.
    Keywords: multi-constraint and multi-objective; tourist routes; personalised recommendation; greedy algorithm.
    DOI: 10.1504/IJRIS.2023.10053462
     
  • Athlete's facial emotion recognition method based on multi physiological information fusion   Order a copy of this article
    by Xingbo Zhou, Junmin Wang 
    Abstract: In order to overcome the problems of low accuracy and high time consumption of athlete’s facial emotion recognition, this paper proposes a method of athlete’s facial emotion recognition based on multi physiological information fusion. First of all, a variety of sensors are used to collect the athlete’s ECG, respiration, pulse and skin conductance. Secondly, wavelet transform is used to process multiple physiological information. Then, the sequential backward floating selection method is selected to delete redundant features. Finally, combining the physiological information features, the least squares support vector machine is used to output the athlete’s facial emotion recognition results. The experimental results show that this method can accurately recognise athletes’ facial emotions. The F1 score of facial emotion recognition is higher than 0.97, and the recognition time is less than 300 ms.
    Keywords: multiple physiological information; information fusion; facial emotion; least squares support vector machine; wavelet transform.
    DOI: 10.1504/IJRIS.2023.10053463
     
  • Complex background image segmentation based on multi-scale features   Order a copy of this article
    by Yanting Cao 
    Abstract: Aiming at the problems of large feature extraction error and poor segmentation effect in complex background image segmentation, a complex background image segmentation algorithm based on multi-scale features is designed. Firstly, the kernel function of multi-scale extraction method is used to initially determine the image feature density, and the Gaussian kernel function is introduced to determine the distance between the feature distribution points and the centre point to complete the image global feature extraction; then, set the grey level constraint of the local feature image to complete the local feature extraction; finally, determine the image edge threshold, divide the complex pixel feature area, determine the image feature membership and fuzzy rate, transform the segmentation problem into a nonlinear problem, and complete the segmentation. The experimental results show that the proposed algorithm can reduce the feature extraction error, the maximum error is less than 1%, and optimises the image segmentation results.
    Keywords: multi-scale feature; complex background image; kernel function; LBF model; optimal curve; image segmentation.
    DOI: 10.1504/IJRIS.2023.10053464
     
  • Personalized Recommendation Algorithm of Music Resources Based on Category Similarity   Order a copy of this article
    by Linqian Peng, Da Li 
    Abstract: Because the collaborative filtering algorithm cannot achieve accurate matching, a personalised music resource recommendation algorithm based on category similarity is proposed. The user preference type parameters are obtained by the modelling and analysis method of knowledge graph, and the personalised preference judgement model is established according to the type parameters. The feature registration algorithm is used to mine the personalised features of music resources, and the homomorphic reliability of the personalised scores is analysed to build a joint parameter matching model of music resources. Finally, through the analysis of category similarity characteristics, the adaptive statistical analysis of music resources is carried out, and the personalised feature parameters of music resources are extracted to achieve personalised recommendation of music resources. The simulation results show that the minimum satisfaction of the method in the optimal state is 92.7%, the resource holding level is always above 92%, and the recommended accuracy is 98.9%, which shows that the method in this paper is more practical.
    Keywords: category similarity; music resources; personalised; recommendation; preference; contribution level.
    DOI: 10.1504/IJRIS.2023.10053465
     
  • An Optimization of Mobile Terminal Data Mining Method Based on Internet of Things   Order a copy of this article
    by Yi Wang  
    Abstract: In this paper, the optimisation of mobile terminal data mining method based on internet of things (IoT) is studied. Firstly, a framework for mobile terminal data mining optimisation is constructed, and mobile terminal data is collected by the mobile agent wireless sensor data acquisition technology. Then the collected data are clustered by the chaotic search particle swarm K-means algorithm, and the clustered data are transmitted to the abnormal access detection module of mobile terminal users. The access detection module finally completes the mining of abnormal access behaviours of mobile terminal users by detecting the abnormal characteristics of user access behaviours, determining the abnormal type and checking the abnormal evolution. The experimental results show that the energy consumption of this method does not exceed 4J in a noisy environment, and this method is low in the data mining energy consumption and high in the accuracy.
    Keywords: internet of things; IoT; mobile terminal; data mining; data acquisition; data clustering; abnormality detection.
    DOI: 10.1504/IJRIS.2023.10053466
     
  • Personalized leisure tourism route recommendation method based on Knowledge Map   Order a copy of this article
    by Chunhua Wang  
    Abstract: In order to improve the efficiency of tourism route recommendation, this paper introduces knowledge map and designs a personalised leisure tourism route recommendation method based on knowledge map. Firstly, build a leisure tourism route knowledge map, generate a tourism route database through tourism notes, and then search the candidate tourism route sequence in the database. Finally, based on the comprehensive analysis of the attributes of tourists such as travel time and user category, determine the score of candidate tourism routes according to the value of tourist attractions, user category and user preference, and take the three routes with the highest score as the recommendation result. The experimental results show that this method can complete the recommendation of tourist routes in about 10 s, and can significantly increase the number of tourists in the scenic spot, with an average increase of nearly 10%, which has certain application value.
    Keywords: knowledge map; personalisation; leisure tourism; route recommendation; data mining; user preference.
    DOI: 10.1504/IJRIS.2023.10053467
     
  • A Recognition method of learning behavior in online classroom based on feature data mining   Order a copy of this article
    by Yingyao Wang  
    Abstract: In order to effectively ensure the recognition effect of online classroom learning behaviour and improve the accuracy and efficiency of online classroom learning behaviour recognition, an online classroom learning behaviour recognition method based on feature data mining is proposed. This paper analyses the concept and process of feature data mining, and extracts the characteristics of learning behaviour data in online classroom. Principal component analysis was used to pre-process the characteristics of learning behaviour data in online classroom. Using the method of feature data mining, this paper constructs the recognition model of learning behaviour in online classroom to realise the recognition of learning behaviour in online classroom. The experimental results show that the proposed method has a good effect on the recognition of learning behaviour in the online classroom, and can effectively improve the accuracy and efficiency of the recognition of learning behaviour in the online classroom.
    Keywords: feature data mining; principal component analysis; online classroom; learning behaviour; behaviour recognition.
    DOI: 10.1504/IJRIS.2023.10053468
     
  • Evaluation of Shortest Path of Network by using an Intuitionistic Pentagonal Fuzzy Number with Interval Values   Order a copy of this article
    by Prasanta Kumar Raut, Siva Behera 
    Abstract: Shortest path problems of weighted graphs and networks are extensively used as optimisation tools in numerous problems in different application domains. This optimisation problem is used for determining the shortest path of a fuzzy associated network between two given nodes known as source and destination. Normally, mathematical models may fail to find the proper result, but fuzzy logic can handle it very easily and give the desired result. This work designs an algorithm that computes a path of minimum length along with the shortest length of an interval-valued intuitionistic fuzzy graph with vertices assuming weights that are crisp numbers and the edges assuming weights that are intuitionistic interval-valued pentagonal fuzzy numbers (IIPFNs). Further, here we explain our result with the help of a numerical problem.
    Keywords: intuitionistic pentagonal fuzzy number; shortest path problem; SPP; ranking method; dynamic programming approach; intuitionistic interval-valued pentagonal fuzzy numbers; IIPFNs.
    DOI: 10.1504/IJRIS.2023.10053470
     
  • A network big data classification method based on decision tree algorithm   Order a copy of this article
    by Nian Xiao, Siguang Dai 
    Abstract: Aiming at the problem of low accuracy and low efficiency of network big data classification, a new network big data classification method based on decision tree algorithm is designed. First, the crawler manager circularly collects network big data, sets the collection threshold and randomly generates crawler signatures, so as to continuously collect and update data; then, construct the directed graph of network big data, automatically select and extract the key feature attributes of network big data, and extract the interference factors of feature data. Finally, the network big data classification decision tree is constructed to obtain the optimal gain data, determine the node attributes of the data, and complete the classification algorithm design combined with recursive call rules and classification termination conditions. Experimental results show that the algorithm can improve the accuracy and efficiency of data classification.
    Keywords: decision tree algorithm; network big data; classification; crawler signature; digraph; information gain; recursive call.
    DOI: 10.1504/IJRIS.2023.10053641
     
  • A method for personalised music recommendation based on emotional multi-label   Order a copy of this article
    by Yuan Luo, Qiuji Chen 
    Abstract: In this paper, a personalised music recommendation method based on emotion multi-label was proposed. First is the analysis of music emotion and music emotional label, then, the principal component analysis method is used to reduce the dimension to process the music features and complete the preprocessing. Secondly, construct the music emotion multi-label, and combine the cosine method to calculate the emotional multi-label similarity. Finally, the interest degree of emotional multi-label is calculated to obtain the user's interest degree of music resources, and the personalised recommendation method is optimised to realise the personalised recommendation of music. Experimental results show that the average coverage rate of personalised music recommendation of the proposed method is as high as 99.5%, the accuracy is 98.3%, and the recommendation time of 500 music items is only 18.9 s. Therefore, the recommendation effect of the proposed method is good, the accuracy of personalised music recommendation is improved, and the recommendation time is shortened.
    Keywords: sentiment multi-label; principal component analysis; TF-IDF method; cosine method; music personalised recommendation.
    DOI: 10.1504/IJRIS.2023.10050585
     
  • A recognition method of basketball players' shooting action based on Gaussian mapping   Order a copy of this article
    by Zhu Xia 
    Abstract: In order to overcome the problem of low recognition accuracy of traditional action recognition methods, this paper proposes a basketball player shooting action recognition method based on Gaussian mapping. Firstly, the basketball shooting image is preprocessed by block initialisation and denoising to improve the quality of the image. Secondly, based on the image preprocessing results, Gaussian mapping is used to extract the target features of shooting action image. Finally, according to the target characteristics, the multi-level feature decomposition and fuzzy processing of the image are carried out to realise the shooting action recognition. Experiments show that the designed method has high accuracy and recall rate, the maximum recognition accuracy reaches 96%, and the recognition time is short, and the number of false recognition frames is less, which shows that the designed method has high practical application performance.
    Keywords: Gaussian mapping; basketball; shooting action; action recognition; feature extraction.
    DOI: 10.1504/IJRIS.2023.10050586
     
  • A rapid recognition of athlete's human posture based on SVM decision tree   Order a copy of this article
    by Nianhui Wang, Qingxue Li 
    Abstract: In order to solve the problems of low recall rate of human posture data collection results, low recognition rate and long recognition time in traditional recognition methods, a rapid recognition method of athlete's human posture based on SVM decision tree was proposed. The Kinect sensor is used to collect the athlete's human posture data, and the mixed Gaussian background modelling method is used to segment the collected athlete's human posture image. Scale normalisation is performed on the segmented images, and a star model is used to extract the pose features of athletes' bodies. According to the characteristics of human posture, the SVM decision tree is used to classify and identify the human posture of athletes. The experimental results show that the maximum recall rate of this method is 98%, the minimum value is 93%, the recognition rate is above 97.2%, and the average recognition time is 0.62.
    Keywords: SVM decision tree; athlete; human posture; rapid recognition; scale normalisation; star model.
    DOI: 10.1504/IJRIS.2023.10050587
     
  • A personalised recommendation method of pop music based on machine learning   Order a copy of this article
    by Honghao Yu 
    Abstract: In order to enhance the satisfaction of pop music personalised recommendation and improve the accuracy and efficiency of pop music personalised recommendation, a pop music personalised recommendation method based on machine learning is proposed. Firstly, the relevant theories of machine learning and short-term and long-term memory artificial neural networks are studied, and then the popular music word vector is extracted by using softmax function, and the collaborative filtering algorithm with weighting factor is introduced to calculate the similarity of popular music word vector. Finally, based on the LSTM network, a pop music personalised recommendation model is constructed to realise pop music personalised recommendation. Experiments show that the method proposed in this paper has a personalised recommendation satisfaction index of 97.8% for pop music, the recommendation time is only 19.4s, and the average value of MAPE and RMSE are only 0.037 and 0.039 respectively. The recommendation accuracy, satisfaction and efficiency are high, and the design purpose can be achieved.
    Keywords: machine learning; Word2Vec; LSTM network; popular music; personalised recommendation.
    DOI: 10.1504/IJRIS.2023.10052087
     
  • Study on weak signal enhancement method in wireless communication under electromagnetic interference environment   Order a copy of this article
    by Cao Chai 
    Abstract: In order to overcome the problems of high bit error rate and high signal enhancement time in weak signal enhancement methods, this paper proposes a weak signal enhancement method in wireless communication under electromagnetic interference environment. Firstly, the transmission characteristics of electromagnetic wave are analysed and the attenuation results of electromagnetic wave signal transmission are determined. Then, the bandwidth of weak signal is determined by band-pass sampling theorem, and the key characteristics of weak signal are determined by time-domain characteristic analysis. Finally, the amplitude modulation coefficient is used to adjust the frequency offset and initial phase through a low-pass filter to enhance the weak signal in wireless communication. The experimental results show that the lowest bit error rate of this method is about 1%, the weak signal enhancement time is no more than 2.6 s, and the weak signal enhancement effect is good.
    Keywords: electromagnetic interference; wireless communication; weak signal enhancement; band-pass sampling theorem; discrete signal; maximum likelihood estimation.
    DOI: 10.1504/IJRIS.2023.10052085
     
  • Research on enterprise financial investment risk prediction method based on binary tree clustering   Order a copy of this article
    by Qian Cao 
    Abstract: In order to overcome the problems of poor accuracy of enterprise financial investment risk prediction and poor fitting effect of operating profit margin, this paper proposes a risk prediction method of enterprise financial investment based on binary tree clustering. Firstly, binary tree clustering is used to classify financial data. Secondly, the binary tree is used to obtain the data probability density function, and then the maximum likelihood estimation method is used to solve the density objective function. Finally, the investment risk prediction results are obtained through the expectation maximisation method to realise the financial investment risk prediction. The experimental results show that the prediction accuracy of this method is as high as 99.06%, and the fitting effect of operating profit margin is good, which shows that this method can improve the prediction accuracy of enterprise financial investment risk.
    Keywords: binary tree clustering; Bayes theorem; covariance matrix; financial investment risk.
    DOI: 10.1504/IJRIS.2023.10052090
     
  • A sleep scheduling algorithm of redundant nodes in power communication sensor networks based on relative local density   Order a copy of this article
    by Binyuan Yan 
    Abstract: In order to overcome the problems of low network coverage and short network node survival time in the traditional node sleep scheduling algorithm, a redundant node sleep scheduling algorithm based on relative local density is proposed in this paper. The composition of power communication sensor network is analysed, and the energy receiving model of redundant nodes in power communication sensor network is constructed. The sleep scheduling of network redundant nodes is realised according to the relative local density redundant node sleep algorithm. The results show that when the number of nodes is 120, the network performance of this method is 5.18 and the network coverage can reach 99.93%. At the same time, the survival time of sensor nodes under this method is longer, this method can prolong the network life cycle, and the sleep scheduling effect of redundant nodes is better.
    Keywords: node sleep scheduling algorithm; energy consumption model; relative local density; network life cycle; relative local density.
    DOI: 10.1504/IJRIS.2023.10051779
     
  • An adaptive fusion method of multi-mode human-computer interaction information in intelligent warehouse   Order a copy of this article
    by Shengbo Sun, Hao Wang, Chong Li, Yi Wang, Bing Li 
    Abstract: In order to obtain comprehensive multi-mode human-computer interaction information and enhance image definition, the adaptive fusion method of multi-mode human-computer interaction information in intelligent warehouse is studied. The canonical correlation analysis (CCA) fusion method of double width learning is established. The multi-mode HMI image information is introduced. Combined with the enhancement characteristics of dimensionality reduction modes at each layer, the enhanced nonlinear fusion dimensionality reduction characteristics are obtained through the fusion node layer, and the adaptive fusion results of multi-mode human-computer interaction information are output at the output layer. The experimental results show that when Gaussian and Poisson noise are added, this method can still adaptively fuse the multi-mode human-computer interaction image information, highlight the image details and improve the image definition. In different dimensions of human-computer interaction image information, it has rich image information and less visual information loss.
    Keywords: intelligent warehouse; multimodal; human-computer interaction information; adaptive fusion; width learning; correlation analysis.
    DOI: 10.1504/IJRIS.2023.10052086
     
  • Application of fuzzy logic in evaluating the authenticity of hadith and narrators   Order a copy of this article
    by Musaab Zarog 
    Abstract: Hadith in Islam is the Arabic word used to represent the traditions and sayings of the prophet. These traditions are normally categorised into four levels depending on their degree of authenticity. This accepted ranking system ignores the categories that might fall between these four ranks. Traditionally, hadith scholars apply a multi-criterion to conclude the degree of authenticity, which mainly depends on their training, experience, and various information reported about the hadith narrators. The research aims to use an arithmetic decision-making approach to judge the degree of validity and degree of authenticity of hadith, considering all factors utilised by hadith scholars such as the reliability and accuracy of the narrator as well as the chain of transmission. The proposed approach will also consider different opinions in the literature about narrators' accuracy, reliability, and memorisation capabilities. The arithmetic method was applied to three hadiths ranking: true, good, and weak, and then the weighted arithmetic mean was calculated for each of them which represent the degree of validity of each hadith. The obtained results agree to a high degree with judgment reported by hadith scholars.
    Keywords: artificial intelligence; natural language processing; decision making; authentication; decision making; Sharia; deep learning; Islamic sciences.
    DOI: 10.1504/IJRIS.2023.10051253
     
  • Study on three-dimensional coverage method for wireless sensor networks based on greedy algorithm   Order a copy of this article
    by Yong-Bo Liu, Yu-Cai Zhou 
    Abstract: In order to reduce the problem of poor network coverage caused by unreasonable stage arrangement, a three-dimensional coverage method based on greedy algorithm for wireless sensor networks is proposed. The node perception model is established to determine the radiation range of a single node; By grid routing method, the target coverage area is divided into equidistant three-dimensional grid structure. The greedy algorithm is used to calculate the maximum distance and the minimum angle of the node location, then the rationality of the node is optimised, and finally the network coverage gap is repaired by edge coverage and node patch, so as to achieve three-dimensional full coverage of wireless sensor network. The experimental results show that the node survival rate of the proposed method is higher than 70%, and the maximum hop number of nodes is 8, and the network 3D coverage effect is better.
    Keywords: greedy algorithm; wireless sensor network; three-dimensional coverage; node perception model; grid routing; overlay.
    DOI: 10.1504/IJRIS.2023.10053469