Please wait a minute...
, No 7

■ Cover page(PDF 3233 KB) ■  Table of Content, July 2023(PDF 89 KB)

  • Are the Customers Receiving Exact Recommendations from the E-Commerce Companies? Towards the Identification of Gray Sheep Users Using Personality Parameters
    Babaljeet Kaur and Shalli Rani
    2023, 19(7): 425-433.  doi:10.23940/ijpe.23.07.p1.425433
    Abstract    PDF (329KB)   
    References | Related Articles
    In the real-world, use of digital media is increasing day by day. Consumers are always in a dilemma when they have a lot of choices and they always get confused among them. The E-commerce companies want to increase their profit, so they introduced recommender systems. The recommender system is like the newspaper which gives us recommendations when we do not have any information about the particular products of the market. Users’ mind diverts the fetching of data from users’ favorite headlines just like the recommender system which fetches the items according to the users’ taste and preferences. However, to date Recommender systems have constraints of cold start problems, gray sheep user problems and shilling attacks. Variety in the tastes of the users gives rise to the gray sheep problem where it is difficult to match the products as per the choice of one person’s taste to the other person. This affects the performance of recommender systems. In this paper, the chocolate bars flavors_of_Cacao dataset (124KB) with 1796 entries tuples and 9 attributes dataset based on the personality parameters (modeling with Big Five model), is analyzed with boosted decision trees, two neural networks, logistic regression and decision forest. The accuracy of the decision forest is validated over other machine learning algorithms for a recommendation of the chocolate bars.
    Suitability Index Prediction for Residential Apartments Through Machine Learning
    Kshitij Kumar Sinha, Manoj Mathur, and Arun Sharma
    2023, 19(7): 434-442.  doi:10.23940/ijpe.23.07.p2.434442
    Abstract    PDF (1028KB)   
    References | Related Articles
    The value a good designed apartment adds to the life of the occupants has largely remained a lived out experience. Additionally, the architectural decision that ensures ‘value’ in designed homes has been restricted to the predetermined housing typologies influenced by both internal and external variables. The lack of exploration to reflect the ‘lived out experience’ as a key internal factor has instigated the continuous and rampant build-rebuild chain, which seems to only end with irreparable environmental degradation. Partnering with Artificial Intelligence to promote mindfulness amongst various stakeholders to address these pressing and urgent concerns has opened numerous possibilities. This paper demonstrates that by mapping the user group (represented through nine family structure scenarios) with their day to day requirements (derived through their indulgence in conducting an activity in terms of time spent and criticality) to help us to assign weightage to the spaces of a residential apartment based on the user group preference (by applying the TOPSIS method). Acting as an extension to the previous works on predicting the usability of a 3BHK apartment, the research presents a strong case of Machine Learning applications to predict the Suitability Index of a residential apartment.
    Recommender System: Towards Identification of Shilling Attacks in Rating System Using Machine Learning Algorithms
    Manpreet Kaur and Shalli Rani
    2023, 19(7): 443-451.  doi:10.23940/ijpe.23.07.p3.443451
    Abstract    PDF (362KB)   
    References | Related Articles
    Recommender System is a software tool that considers multiple parameters like the interest of a user, historical user data, browsing behavior, etc. to provide suitable recommendations. From e-commerce to social websites and from online research articles to entertainment, recommender systems have successfully rooted themselves in almost every application domain. Due to the popularity gain and people reliance on the recommendations generated in almost every domain, recommender systems have become vulnerable to online attacks known as shilling attacks or false rating attacks. An insight of recommender systems along with an analysis of shilling attacks is performed in this paper. These attacks are very common in social media to make the content popular among users by false ratings. Therefore, to find out the effect of false ratings, analysis of average, bandwagon attack, hybrid, and segment attacks are done on the basis of specificity, sensitivity, recall, precision and F-measure on Netflix and Movie lens dataset. These attacks should not affect food product items because most people purchase online items on the basis of ratings only. A good recommender system can overcome these attacks. Therefore, to deal with the above challenge of false ratings’ attack, an ensemble learning model has been proposed to classify the individuals which are most similar to the target user so that relevant recommendations are made to each user as well. The proposed approach is validated for dataset of the German online reviews/rating of organic coffee, to determine the customer demand for a particular brand of coffee using Machine learning algorithms. The comparative results show that proposed approach is better than Locally Deep Support Vector Machine,Neural Network, Logistic Regression, Bayes Point, and Decision Forest algorithms with an accuracy of 98 percent.
    Exploratory Review of Machine Learning-Based Software Component Reusability Prediction
    Srishti Bhugra and Puneet Goswami
    2023, 19(7): 452-461.  doi:10.23940/ijpe.23.07.p4.452461
    Abstract    PDF (1518KB)   
    References | Related Articles
    Software reusability is recognized as a crucial aspect of quality. Addressing the software turmoil, increasing software quality, and enhancing performance are the most evident benefits of software reuse. Finding reusable software elements in an identified present structure is a critical yet underdeveloped challenge. Authors employ a method built on software models and metrics to discover and assess reusable software. This investigation aims to evaluate the efficacy and competency of machine learning methods that are being used to create an accurate and useful assessment framework that can evaluate the reusability of software elements using static metrics. In the present work, the authors conduct a thorough literature review of machine learning methods used to forecast software reusability. Initially, background information and relevant studies are presented. After that, a summary of machine learning techniques is provided. Additional research is being done to assess how well different machine learning methods forecast software reusability. The highest-scoring ML classification framework achieved an accuracy of 89.33% (ANN), outperforming other studies in predicting accuracy (e.g., KNN, DT, RF, SVM, BT, KNN, and HMM). The outcomes may be employed to determine which machine learning model is most useful for identifying reusable parts of software since it is accurate, quick, productive, and cost-effective.
    Analyzing Brain Signals for Predicting Students’ Understanding of Online Learning: A Machine Learning Approach
    Harsha Gaikwad, Sanil Gandhi, Arvind Kiwelekar, and Manjushree Laddha
    2023, 19(7): 462-470.  doi:10.23940/ijpe.23.07.p5.462470
    Abstract    PDF (424KB)   
    References | Related Articles
    The primary focus of the educational process is on students who must comprehend the course material. Various methods predict a student's understanding, including questioning, exams, quizzes, feedback, and observing facial expressions. In an offline teaching-learning process, it is relatively straightforward to predict students' understanding. However, online learning, particularly while watching videos, presents challenges due to distractions in the surrounding environment. Predicting a student's understanding level in online learning becomes tedious as there is limited personal interaction between the student and the teacher. This study aims to identify an optimal machine learning model for predicting students' understanding of online learning by analyzing brain signals recorded using Electroencephalogram (EEG). The dataset comprises of brain signals collected from students of different ages and educational backgrounds. The GridsearchCV technique is utilized to select the optimal parameters. The experimental results demonstrate that KNN and SVM achieve nearly identical accuracy, approximately 99%, for predicting the understanding level.
    EDocDeDup: Electronic Document Data Deduplication Towards Storage Optimization
    Me Me Khaing and N. Jeyanthi
    2023, 19(7): 471-480.  doi:10.23940/ijpe.23.07.p6.471480
    Abstract    PDF (875KB)   
    References | Related Articles
    Understanding data deduplication in storage is essential for investigating the optimization of various data storage issues. For detecting and removing duplicate data, data deduplication has become an important and cost-effective optimization technique. Storage issues in the storage area for organizations exist, and if not conveyed to optimize there and then, a slower rate of storage capacity is expected. The proposed system (EdocDedup) addresses the aforementioned issue by applying data deduplication technique and implementing SHA-256 for hash value calculation and only keeping the unique hash values on an electronic document’s dataset containing word files, text files, html files, excel files, zip files, pdf files, and PowerPoint presentation files. By demanding the proposed technique there is a benefit in storage saved and a variety of duplicate files are explored efficiently. EdocDedup's performance is achieved through the use of user-uploaded files.
    Multi-Objective Optimization of Cancer Treatment Using the Multi-Objective Grey Wolf Optimizer (MOGWO)
    Linkai Chen, Honghui Fan, and Hongjin Zhu
    2023, 19(7): 481-490.  doi:10.23940/ijpe.23.07.p7.481490
    Abstract    PDF (542KB)   
    References | Related Articles
    The use of mathematical modeling to study biological phenomena is one of the best methods available for studying these phenomena. The development of mathematical models for simulating, controlling, and predicting phenomena has always been significant. The application of mathematical models in optimization is one of the advantages of using them. In the context of cancer treatment, the goal is to reduce the concentration of cancer cells during the treatment period through optimal control. An important issue that was not considered in previous studies was the concentration of the active drug, which had a significant influence on the clinical health of the patients. The aim of the current study was to establish a protocol for optimal drug administration by minimizing the concentration of cancer cells and the concentration of the drug. The multi-objective grey wolf optimizer (MOGWO) algorithm was used for the first time to solve this multi-objective problem and the results were compared to those obtained with the NSGA-II algorithm.
ISSN 0973-1318