International Journal of Intelligent Systems Technologies and Applications (6 papers in press)

Regular Issues

• A novel framework for labelling Duplicate and Non-Duplicate bugs  
  by Kulbhushan Bansal, Sunesh Sunesh, Manju Rohil, Harish Rohil 
  Abstract: Bug handling is an essential part in the software development life cycle. It can be very cumbersome, tedious and error-prone due to the complexity and size of software projects and teams. Duplicate bugs make the bug handling process even more tedious. In this paper, binary duplicate detection and ranking-based duplicate detection mechanisms have been combined together to deal with a two way duplication mechanisms. A novel framework has been proposed which predicts the label (duplicate or non-duplicate) for any newly arrived bug report. Further, if found as duplicate, the proposed framework produces a ranked list of bug reports which might be similar to the duplicate predicted bug report. The proposed framework has been experimentally validated using bug reports obtained from Eclipse, NetBeans and Mozilla Firefox projects of Bugzilla repository. From the experimental evaluations, we observed that deep learning-based models outperform traditional machine learning algorithms in bug report classification.
  Keywords: duplicate bug detection; information retrieval; machine learning; classification; deep learning; rank aggregation; mining software repositories.
  DOI: 10.1504/IJISTA.2023.10055781
   
  
• Modeling and simulation of cancelling orders behavior by driver and passenger on the ride-hailing platform  
  by Wei Zhang, Ruichun He, Yong Chen 
  Abstract: It is common to cancel orders by driver and passenger on the ride-hailing platform, which will reduce operational efficiency. This paper studies the behaviour of cancelling orders for ride-hailing platform. Firstly, a model of cancelling orders decision for driver and passenger based on prospect theory is built. Secondly, a simulation algorithm of vehicle carrying passengers with the behaviour of cancelling orders is proposed. Thirdly, simulation is carried out and relative results are obtained. Finally, some strategies and suggestions for ride-hailing platform are given. The results show that cancelling orders behaviour influences driver income and platform profit. Cancelling order rate of driver and that of passenger are influenced by various factors respectively. The cancelling order rate can be decreased by adjusting the factors, such as exoneration time, subsidy, penalty value, etc. These conclusions would help us to developing relative strategies, which has certain guiding significance for management of ride-hailing platform.
  Keywords: cancelling orders; ride-hailing platform; prospect theory; simulation; multi-agent.
  DOI: 10.1504/IJISTA.2023.10055916
   
  
• ITTS Model: Speech Generation for Image Captioning using Feature Extraction for End-to-End Synthesis  
  by Tushar Ghorpade, Subhash K. Shinde 
  Abstract: The current growths in e-content, information exchanged through social media, e-news, etc. Several researchers have proposed an encoder-decoder model with impressive accuracy. This paper exploits feature extraction from images and text for the encoder model using a word embedding method with proposed convolutional layers. State-of-the-art image-to-text and text-to-speech (ITTS) systems learn models separately, one describes the content of an image and the other follows with speech generation. We adopted the Tacotron model for the naturalness of a text with most popular datasets. It can also consistently analyse using evaluation metrics like bilingual evaluation understudy (BLEU), METOr, and mean opinion scale (MOS). The proposed method can significantly enhance the performance and competitive results of a standard image caption and speech generation model. The results show that we obtained an improvement by almost 4% to 5% BLEU score in image captioning model and approximately MOS is 3.73 in speech model.
  Keywords: image captioning; convolutional neural network; recurrent neural network; RNN; sequence to sequence language model; text-to-speech synthesis model; Tacotron model; LSTM.
  DOI: 10.1504/IJISTA.2023.10056176
   
  
• Differential Evolution Variants for Optic Disc Localisation in Eye Fundus Images Using Entropy Measure  
  by Chen Zhang, B.Vinoth Kumar, Siwen Zhang, J. Prakash, Shiman Wen, Bineeshia Joel, Wenjin Li 
  Abstract: Retinal images are broadly used to expose several retinopathies like glaucoma, optic neuritis, macular degeneration related to age, and diabetic retinopathy. Computer-aided retinal image diagnoses efficiently facilitate clinicians by automating the load screening process to diagnose these conditions of which optic disc (OD) segmentation is the primary phase and categorised as a critical search problem. The proposed approach incorporates the preliminary processing of retinal fundus images and optical disc localisation using three variants of differential evolution (DE) for this application: standard differential evolution (SDE), cellular differential evolution (CDE), and unified differential evolution (UDE). To localise optic disc effectively, Entropy of an image region is proposed as a fitness function. The performance of the suggested technique is examined on two datasets: DIARETDB0 and DIARETDB1. The results suggested that the proposed methodology has achieved a 99.23% detection accuracy and demonstrates the dominance of localising the optic disc over the others.
  Keywords: differential evolution; optic disc localisation; entropy; fitness function; fundus image; diabetic retinopathy; glaucoma.
  DOI: 10.1504/IJISTA.2023.10056272
   
  
• Automated Poetry Scoring Using BERT with Multi-Scale Poetry Representation  
  by Mingzhi Gao, Selin Ahipasaoglu, Kristin Schuster 
  Abstract: Automated poetry scoring is an emerging task in automated text scoring, which is receiving increasing attention in AI for education. Poetry is distinct from other text in its complexity and specialty in language feature moreover, poems are usually rated from multiple criteria besides the overall impression. However, few existing methods to the best of our knowledge have considered a tailored text representation model for encoding poetry. Moreover, the lack of large poetry corpus and extensive labelled data is another major constraint to construct an effective poetry scoring model. To address such problems, we proposed BERT-based models with multi-scale poetry representation. In addition, we employ multiple losses and R-Drop strategy to align the distribution of manual and model scoring and mitigate the tendency of consistent score in poems. Experiment results demonstrate that our model with multi-scale poetry representation stands out when comparing with single-scale representation model.
  Keywords: automated poetry scoring; pre-trained language model; multi-scale text representation.
  DOI: 10.1504/IJISTA.2023.10056521
   
  
• Jitter as a quantitative indicator of dysphonia in Parkinson's disease  
  by Jennifer Saldanha, Malini Suvarna, Dayakshini S, Cynthia Santhmayor 
  Abstract: A non-invasive way of diagnosing Parkinson's disease from speech signals is presented in this paper. A variety of frequency, amplitude, harmonicitynoise, and cepstral features are extracted from speech samples, resulting in a feature vector of 82 coefficients. k-nearest neighbours (k-NN) with k = 10 and artificial neural network (ANN) are applied to the dataset on individual and combined features to detect Parkinson's disease. The jitter feature obtained a maximum accuracy with both k-NN and ANN classifiers. k-NN outperformed ANN by obtaining a classification accuracy of 90% for jitter local features and 88.3% for combined features. The severity of the disease is assessed using multi-class classification, obtaining an overall accuracy of 83.6% and 82.4% for k-NN and ANN, respectively. The accuracy in detection is also verified on the dataset divided based on age and gender category. The results of the perceptual test proved that the predominant voice quality in Parkinson's disease is hoarse.
  Keywords: Parkinson's disease; mel frequency cepstral coefficients; linear prediction cepstral coefficients; k-nearest neighbour; k-NN; multi-layer perceptron artificial neural network.
  DOI: 10.1504/IJISTA.2023.10056658