International Journal of Information Technology and Management (13 papers in press)

Regular Issues

• Multilevel secure storage method of electronic documents based on Hash function  
  by Runhua Miao 
  Abstract: In order to improve the storage security and efficiency of electronic documents, a new multi-level secure storage method of electronic documents based on Hash function is proposed in this paper. Firstly, the vector space model is constructed, and the electronic document data is divided into fixed size data blocks by using the idea of linear segmentation. Then, after normalisation, Hash function is used to encrypt each data block. Finally, according to the encryption results, the compressed sensing method is used for multi-level secure storage of electronic documents. The experimental results show that, compared with the traditional storage methods, this method has strong encryption performance, and the maximum intrusion rate is no more than 0.5%, which can ensure the security of electronic documents. This method can improve the storage efficiency of electronic documents.
  Keywords: Hash function; electronic document; multi-level security storage; data segmentation.
  DOI: 10.1504/IJITM.2022.10051753
   
  
• An Enterprise financial credit risk measurement method based on differential evolution algorithm  
  by Lixia Du, Xin An 
  Abstract: In order to reduce the time cost and risk misjudgement rate of financial information risk measurement, this paper proposes a new enterprise financial credit risk measurement method based on differential evolution algorithm. Firstly, after preprocessing the enterprise financial credit risk data and determining the location of the clustering centre, a differential evolution automatic clustering model is constructed. Secondly, according to the clustering results, the differential evolution algorithm is used to measure the basic process of enterprise financial credit risk. Finally, the improved differential evolution algorithm is used for iterative measurement to achieve enterprise financial credit risk data measurement. The experimental results show that the time cost of the proposed method for enterprise financial credit risk measurement can be controlled within 0.4 s, and the error rate is not more than 1% under the condition of 1,000 data.
  Keywords: differential evolution algorithm; corporate finance; credit risks; measurement method.
  DOI: 10.1504/IJITM.2022.10051754
   
  
• Performance evaluation method of human resource management based on Chaotic Algorithm.  
  by Lina Si, Zhanlei Shang 
  Abstract: This paper proposes a new method of HR management performance evaluation based on chaotic algorithm. First of all, to improve the evaluation performance as the research goal, design the guiding principles of evaluation. Second, under the guidance of the principle, the multi type data of HR management are collected, and the chaotic algorithm is used to extract the characteristics of the data. Finally, after constructing the index system, calculate the weight and construct the evaluation function to complete the overall management performance evaluation. The test results show that the evaluation performance of the method has been improved. On the basis of shortening the evaluation time, the evaluation accuracy has been improved to 97.68%.
  Keywords: chaotic algorithm; human resources; management performance evaluation; index system.
  DOI: 10.1504/IJITM.2022.10051755
   
  
• A data integrity detection method for accounting informatization based on homomorphic hash function  
  by Zhao Guang, Zhi Li 
  Abstract: In order to solve the problems of low data detection accuracy and high detection time overhead, this paper proposes an accounting information data integrity detection method based on homomorphic hash function. First, the accounting data is collected by data mining method and the strong relevance of the data is determined by association rules. Then, set the distance matrix to determine the data key points, match the niche factor between the data key points, and complete the feature extraction. Finally, the binary code is used to mark the accounting information data, and the anti-collision of homomorphic hash function is used to complete the projection of accounting data, so as to realise the data integrity detection. The results show that the detection accuracy of this method is up to 98%, and the detection time overhead is within 4S, which shows that this method can effectively improve the integrity detection effect.
  Keywords: homomorphic hash function; accounting informatisation; data detection; integrity: association rules.
  DOI: 10.1504/IJITM.2022.10051756
   
  
• New business management model of enterprises based on data-driven  
  by Xiaofeng Zhang 
  Abstract: In the current enterprise management mode, there are problems such as low efficiency of enterprise management data processing and reducing the economic benefits of enterprises, which affect the rapid development of enterprises. In order to solve this problem, this paper studies the new business management model of enterprises based on data-driven. Build a data-driven enterprise management mode framework, integrate enterprise management data with KNN algorithm, and calculate user access trust and reliability values with trust management model to improve data processing efficiency and data security. Based on the digital processing of enterprise management data, the development strategy of new business management mode is given. The experimental results show that after applying the management mode designed in this paper, the maximum profit of the enterprise can reach 20.5 million yuan, and the maximum value of the enterprise data processing time is only 6.03 s, which proves that the designed management mode is more efficient for the enterprise management data processing, and can effectively improve the enterprise economic income, and has certain practical application value.
  Keywords: data-driven; business management model; data integration; KNN algorithm; trust management model; development strategy.
  DOI: 10.1504/IJITM.2022.10051929
   
  
• A Recognition method of abnormal learning behavior in MOOC online education based on background subtraction  
  by Hongmei Wan 
  Abstract: In order to overcome the problems of high time-consuming and poor recognition accuracy of learning behaviour recognition, this paper proposes an abnormal learning behaviour recognition method for MOOC online education based on background subtraction. Firstly, the characteristics of students’ abnormal learning behaviours are collected and extracted. Then, the background difference algorithm is used to obtain the foreground object and background of the learning image, and the image pixels are classified. Finally, the mean background method is used to obtain the learning background, the abnormal behaviour recognition classifier is designed, and the background subtraction method is used to realise the abnormal learning behaviour recognition. The results show that the recognition accuracy of this method is as high as 98.32%, the recognition time is only 0.52 s, and the recognition recall rate is as high as 96.7%, indicating that this method can improve the recognition effect of abnormal learning behaviour.
  Keywords: background subtraction; binarisation treatment; mean background method; background difference method; online education.
  DOI: 10.1504/IJITM.2022.10051930
   
  
• An evaluation method of government digital service quality based on big data  
  by Siguang Dai, Zhongming Tang, Ling Zhou 
  Abstract: Aiming at the problems of low accuracy and poor reliability of government service quality evaluation, this paper proposes a government digital service quality evaluation method based on big data. Firstly, the tree structure of index selection was obtained by big data analysis method, and the service quality evaluation index was selected. Secondly, the factor analysis method is used to analyse the reasonability of each evaluation factor and construct the evaluation system. Secondly, it obtains the government digital service data, and finally calculates the weight of service quality evaluation index. According to the clustering of big data, the fuzzy language is integrated and processed to realise the evaluation of government digital service quality. The results show that the evaluation time of the proposed method is not more than 29 s, and the consistency coefficient value of quality evaluation accuracy is 0.79. It has high evaluation efficiency and accuracy.
  Keywords: big data; government digital service; quality evaluation; evaluation index; fuzzy logic.
  DOI: 10.1504/IJITM.2022.10051931
   
  
• A Performance evaluation method of new business model based on grey correlation algorithm  
  by Yan Wang 
  Abstract: A new business model performance evaluation method based on grey correlation algorithm is designed to solve the problems of large evaluation error and low key of screening indicators in the new business model performance evaluation. First, analyse the new business model and screen the performance evaluation indicators of the new business model; then, the clustering algorithm is used to determine the cluster family of each index, extract the performance evaluation index characteristics of the new business model, and construct the performance evaluation index system. Finally, the grey correlation algorithm is used to determine the grey correlation degree between the indicators, quantify the evaluation indicators, build a grey correlation model for the performance evaluation of the new business model, and realise the performance evaluation. The experimental results show that the evaluation error of the proposed evaluation method is only 2%, and the key degree of the selected index is higher than 90%, which is increased by more than 5%. This method has higher practical application value.
  Keywords: grey correlation algorithm; new business model; performance evaluation; clustering algorithm; main sequence; correlation sequence; consistency check.
  DOI: 10.1504/IJITM.2022.10052229
   
  
• Research on digital English teaching materials recommendation based on improved machine learning  
  by Miao Ma 
  Abstract: In order to overcome the problems of low accuracy, time-consuming and low user satisfaction in traditional methods, a digital English teaching materials recommendation method based on improved machine learning is proposed. Firstly, use web crawlers to obtain the data of digital English teaching platform, and use Word2vec model data for training to obtain the data feature vector. Secondly, K-means algorithm is used to cluster users according to feature vectors, and multi-Markov chains are used to predict user interest. Finally, the decision tree algorithm in machine learning is improved on the gradient boosting framework, and the digital English teaching materials are recommended by using the improved algorithm and the user interest prediction results. The experimental results show that the accuracy of this method is more than 96%, the average time of digital English teaching materials recommendation is 76.1 ms, and the average user satisfaction is 96.6.
  Keywords: improved machine learning; digitisation; English teaching; data recommendation; multiple Markov chains; decision tree.
  DOI: 10.1504/IJITM.2022.10052230
   
  
• The personalized classification of brand promotion information based on K-means algorithm  
  by Xi Li  
  Abstract: In order to improve the efficiency of personalised classification of brand promotion information and shorten the time of personalised classification, this paper proposes a personalised classification method of brand promotion information based on K-means algorithm. First, collect brand promotion information, and calculate the text relevance of brand promotion information through weighting factors; Secondly, the attribute division of extension information is carried out by using the three branch decision-making theory; then, the information features of brand promotion are extracted by capsule network; Finally, calculate the similarity between different brand promotion information, obtain the brand promotion information classification function, and realise the personalised classification of brand promotion information through k-means algorithm. The experimental results show that the classification accuracy of this method is 98.08%, and the time of personalised information classification is only 1.20 s, indicating that this method can effectively improve the efficiency of personalised classification of brand promotion information.
  Keywords: k-means algorithm; information attribute division; feature extraction; personalised classification of information.
  DOI: 10.1504/IJITM.2022.10052258
   
  
• Users' satisfaction evaluation based on ISO standards for tourism and travel mobile applications  
  by Carolina Almeida, Bráulio Alturas 
  Abstract: This study is based on the results of a survey applied to a sample of 201 users which gave them an opportunity to identify points of improvement that will be used in a proposal for a new application. In addition, it aimed to obtain data that allows analysis metrics to be applied to international standard organisation (ISO) standards and attempts to understand why individuals use these applications, based on the technology acceptance model (TAM). From the results obtained, it was concluded that there was an average level of satisfaction with the applications of 4.22, on a scale ranging from 1 to 5. Based on the opinions of the respondents, it was also possible to select future improvements for the applications that could contribute to a new tourism and travel application. Finally, it was concluded that there is a direct link between content acceptance and quality.
  Keywords: application; user satisfaction; travel; tourism; rating; TAM; ISO standards; mobile applications.
  DOI: 10.1504/IJITM.2023.10062400
   
  
• BLOCKCHAIN GOVERNANCE: REDUCING TRUSTED THIRD PARTIES WITH DECRED PROJECT  
  by Marcelo Martins, Pedro Campos, Isabel Mota 
  Abstract: Decred is a cryptocurrency with its own blockchain and has several similarities with Bitcoin but implements a governance model that resembles a company with thousands of investors. These stakeholders invest their coins, get the right to direct the project as they see fit and are rewarded for doing so. This research investigates how Decred Project created its own version of money and implemented security measures to improve governance and remove trusted third parties from money issuance and e-voting. This topic is particularly relevant to understand how blockchain technologies improve governance and avoid the tyranny of the majority. In order to reach our goal, we use multi agent simulation and statistical modeling to verify in what extent Decred is capable of providing a predictable, scarce, trustworthy digital asset. We show that Decred increased blockchain security with its hybrid PoW+PoS security mechanism, making an attack more expensive.
  Keywords: governance; blockchain; decred; bitcoin; security; e-voting; proof-of-work (POW); proof-of-stake (POS); consensus; money; cryptoassets; cryptocurrencies.
  
  

Special Issue on: LISS 2017 Emerging Trends, Issues and Challenges in Big Data and Its Implementation

• Evolution of intellectual structure of data mining research based on keywords  
  by Yue Huang 
  Abstract: Data mining has made rapid progress in the past decade and detecting intellectual structure of data mining research is of great help to researchers. We retrieved 5380 papers, published in 11 leading journals of data mining defined by Google Scholar, from SCIE under Web of Science and Scopus databases between 2007 and 2016 to carry out bibliometric analysis. As indicated by the analysis on the evolution of keyword frequency, the research focus of data mining has shifted from such topics as association rule mining to large-scale complex networks. Matrices of high-frequency keywords were also built for different time periods, namely 2007 to 2016 for the whole picture during these years, 2007 to 2011 and 2012 to 2016 for two periods. Clustering results show there are four main data mining topics and the attention has been paid more to graph data mining and complex network analysis in the past 5 years.
  Keywords: data mining; intellectual structure; co-word analysis; clustering; evolution analysis.