International Journal of Networking and Virtual Organisations (24 papers in press)

Regular Issues

• The Nexus Between Allied Policies of GST And FDI With Dependent Telecom Policies of Licensing and Universal Service in India.  
  by Pankaj Mishra, Netra Pal Singh, Ayesha Farooq 
  Abstract: The Indian telecom industry has witnessed sustained grievances from the operators on telecom and allied policies. This paper analyses if such behavior is due to gaps in policy formulation, gaps in implementation or are an ex-post rent seeking behavior. Using the conceptual framework from Lemke and Harris-Wai (2015), the author analyzed if there was adequate telecom stakeholder participation in FDI and GST policy formulation. The causal map of regulation from Coglianese (2012), was used to evaluate gaps in the implementation of FDI and GST for telecom. In FDI policy formulation, the stakeholders were included at the discretion of the DPIIT , and the process is mainly inter-ministerial. In GST policy formulation the operators and DoT were involved in all the stages. DoT did all required changes in the telecom policy that were triggered by FDI and GST policies. Stakeholder participation did not result in lower ex-post issues rather the provisions
  Keywords: Policy Formulation; Stakeholder Participation; rent seeking; telecommunications.
  DOI: 10.1504/IJNVO.2023.10054116
   
  
• The Triggers on Compulsive Online Shopping of Jeans  
  by D. Manimegalai, Senthilkumar S 
  Abstract: This study investigates the antecedents of compulsive online shopping for jeans and identifies the compulsive shoppers of jeans. This descriptive research collects data through an online survey; the sample size is 205
  Keywords: compulsive online shopping; internal and external triggers; online usage; buzz.
  DOI: 10.1504/IJNVO.2023.10055603
   
  
• An Efficient Optimal Load Balancing Algorithm For Distributed File Systems in Cloud Environment  
  by Manjula H. Nebagiri, Latha P. H 
  Abstract: Efficient operations in distributed environments can be obtained by load balancing (LB). LB has turned out to be a vital and interesting research area with respect to the cloud owing to the swift augmentation of cloud computing, and the more services together with better outcomes demand of the clients. The work has developed a framework named an efficient optimal LB (EOLB) for distributed files system to beat the challenges faced in LB. LB was done by means of the framework centred on node distribution together with task distribution. Centred upon the data aspects as well as cloud servers, say CPU in addition to memory usage, together with disk IO occupancy rate, etc. it renders task distribution. Experimental analysis exhibits that the framework attains a better response rate of 74.68 ms, and a processing time (PT) of 0.43 ms, in addition, remains to be efficient when weighed with the prevailing methods.
  Keywords: cloud computing; virtual environment; distributed cloud computing; load balancing; improved K-means clustering; IKMC; modified cockroach swarm optimisation; MCSO.
  DOI: 10.1504/IJNVO.2023.10056519
   
  
• Architectural Framework For Multiplayer Cooperative Cloud Gaming To Optimize Quality Of Service  
  by Nimmagadda Srilakshmi, Naresh Kannan 
  Abstract: The gaming industry is getting more attraction from cloud services providing gaming applications for cooperative multiplayer gaming. Real-time services like cloud gaming are possible by performing necessary process-intensive tasks within the cloud. In this paper, an architectural model for supporting cooperative gaming towards multiplayer is proposed to improve the quality of service in terms of bandwidth and latency compared to existing architectural models. According to this study, the time it takes to change a video has grown by 95% compared to how well Ad hoc mobile cloudlets and the cloud do at sharing videos. Although P-frames, I-frames, and B-frames account for 23%, 78%, and 94% of the resource consumption, density in intra-stream P-frames is also considered. Similarly, resource and CPU use based on skewness are effective compared to mobile devices’ 95% faster video sharing and 77% shorter link delays.
  Keywords: cloud gaming; quality of service; cooperative multiplayer gaming; social networking.
  DOI: 10.1504/IJNVO.2023.10056762
   
  
• Research on consumer satisfaction prediction of e-commerce social platforms based on deep transfer learning  
  by Hehua Mao 
  Abstract: Consumer satisfaction prediction can help e-commerce social platforms adjust their marketing plans, so this article proposes a consumer satisfaction prediction method of e-commerce social platforms based on deep transfer learning. Firstly, product reviews were collected from consumer on e-commerce social platforms, the value of each keyword through the TF-IDF method was calculated, and product characteristics were determined. Then, consumer emotions are divided, and a deep transfer learning model is used to eliminate false comments. Finally, taking the overall emotional value of each comment as a dependent variable and the characteristic emotional value as an independent variable, a multiple regression model for predicting consumer satisfaction is established to obtain relevant results. The experimental results show that the accuracy of the segmentation of consumer sentiment based on this method can reach 96.02%, and the accuracy of consumer satisfaction prediction can reach 0.90. The prediction effect is good.
  Keywords: deep transfer learning; e-commerce social platforms; consumer; satisfaction prediction; TF-IDF method; multiple regression model.
  DOI: 10.1504/IJNVO.2023.10059262
   
  
• User Consumption Behavior Prediction Method in the Context of Social Media Marketing  
  by Gang Chen, Yixi Zhang 
  Abstract: In order to improve the accuracy of user consumption behaviour prediction, mining accuracy of behaviour data, and reduce the long prediction time, this paper designs a user consumption behaviour prediction method in the context of social media marketing. Firstly, research on social media marketing platforms is conducted, and then the impact model between social media marketing and user consumption behaviour is analysed, in order to analyse the impact of social media marketing on user consumption behaviour. Then the user consumption behaviour data is obtained and deeply mined. With the help of integrated weak learning classifier, the behaviour data is classified, and the information entropy of different behaviour data is calculated. Finally, a prediction model based on random forest is constructed to achieve the final prediction analysis. The test results show that the proposed prediction method reduces the prediction error, has high mining accuracy and short prediction time cost for behavioural data, and has good application value.
  Keywords: social media marketing; consumer behaviour; prediction method; impact model; weak learning classifier; random forest.
  DOI: 10.1504/IJNVO.2023.10059263
   
  
• Electronic commerce information personalized recommendation method based on social network data mining  
  by Ping Wen, Ding Ding 
  Abstract: In order to improve the accuracy of e-commerce platform recommendations for users and shorten the recommendation time, this study conducted a personalised recommendation method for e-commerce information based on social network data mining. First, use the crawler algorithm to complete social data mining, extract data features to build user profiles, then use the collaborative filtering algorithm to complete the extraction of user preference features, use the Pearson algorithm to calculate the similarity between the products that users browse for a long time and the user preference features, finally complete the matching calculation between the preference features and the information of the products to be recommended, and then complete the output of the recommendation set. The experiment proves the progressiveness of the proposed method. The results show that the recommendation accuracy of this method is higher than 90%, and the recommendation time is within 15 s. This method can effectively improve the recommendation accuracy and reduce the time cost of recommendation, and has great application value.
  Keywords: data mining; electronic commerce; personalised recommendation; user portrait; preference features.
  DOI: 10.1504/IJNVO.2023.10059266
   
  
• The feature classification method of mobile e-commerce big data under the webcast mode  
  by Jie Li  
  Abstract: In order to improve the special effect of classification results and reduce the convergence classification tolerance, a feature classification method of mobile e-commerce big data under the webcast mode was designed in this paper. Firstly, the dynamic mining of mobile e-commerce big data is realised by setting sliding windows and according to the association mining rules between data. Then, the data weight is calculated based on the TF-IDF method, and the data features are extracted through normalisation processing. Finally, after the data features are derived, the multi-label merging is implemented, and then the feature classification is realised by traversing the feature values using the fuzzy mathematics theory. Experimental results show that the classification specificity interval of this method is [93.1%, 95.6%], and the convergence classification tolerance interval is [0.029, 0.055], indicating that the reliability of this method is high.
  Keywords: webcast mode; mobile e-commerce data; dynamic mining; fuzzy mathematics; feature classification.
  DOI: 10.1504/IJNVO.2023.10059267
   
  
• Consumer credit evaluation of mobile e-commerce platform based on random forest  
  by Ming Yang  
  Abstract: With the goal of improving recall rate and reducing information entropy of evaluation index, a mobile e-commerce platform consumer credit evaluation method based on random forest is designed. First, we use the honeycomb algorithm to optimise the size and depth of the decision tree and the capacity of the feature subset of the classification attribute, and use the random forest to mine the classification of consumer data. Then, the evaluation index system is constructed, and the index weight value is calculated according to the information entropy. Finally, establish a comment set and determine the membership degree of the indicators, and combine the weight of the indicators with the judgment matrix to obtain the consumer credit evaluation results. Experiments show that the recall rate of this method is between 83.74% 85.28%, the information entropy is always lower than 0.02, and the maximum AUC value can reach 0.928.
  Keywords: mobile e-commerce platform; consumer credit; credit evaluation; random forest algorithm; decision tree.
  DOI: 10.1504/IJNVO.2023.10059268
   
  
• Analysis model of short-term search behavior guidance of e-commerce platform users based on knowledge graph  
  by Bin Li, Zhisheng Zhou 
  Abstract: In order to solve the problems of poor satisfaction with product guidance and low user order rate in existing e-commerce platforms, a knowledge graph based short-term search behaviour guidance analysis model for e-commerce platform users is proposed. Firstly, construct a knowledge graph to collect samples of users' short-term search behaviour; Then, attribute preference weighting is applied to users' short-term search behaviour, and the optimal sequential search theory is introduced to construct a user short-term search behaviour guidance analysis model; Finally, matrix decomposition method is used to extract the features of users and products, achieving short-term search behaviour guidance analysis for users. The results show that after using the guidance analysis method in this article, user satisfaction with the product and user order rate can always reach over 90%, indicating good application performance.
  Keywords: knowledge graph; optimal sequential search strategy; attribute preference weighting; search behaviour; matrix decomposition.
  DOI: 10.1504/IJNVO.2023.10059269
   
  
• The Impact of Social Media on Consumer Purchase Intention on E-commerce Platforms  
  by Yuanyuan Niu 
  Abstract: Studying the impact of social media on the purchase intention of consumers on e-commerce platforms not only contributes to the rapid development of e-commerce platforms, but also improves consumer satisfaction. Therefore, the impact of social media on consumer purchase intention on e-commerce platforms was studied. Relevant scales were designed and questionnaires were distributed to different respondents. Validity and reliability tests were conducted on the questionnaire results, and the data that passed the tests were taken as sample data. LLE method was used to reduce the dimension of data, and the influence model of social media on consumers’ purchase intention was built. The experimental analysis results show that the usefulness of search, exchange of comments, forwarding and sharing, information value, and information professionalism have a significant positive impact on consumers’ purchase intention. After positive intervention on social media, the proportion of consumers with purchase intention and satisfaction are higher.
  Keywords: social media; e-commerce platforms; consumer purchase intention; scales; questionnaires; LLE method.
  DOI: 10.1504/IJNVO.2023.10059331
   
  
• Customer Interest Classification Method of E-commerce Trading Platform Based on Decision Tree Algorithm  
  by Xiaowei Ma, Xin Yao, Shizhong Guo 
  Abstract: Due to the large number of customers and diverse interest characteristics in e-commerce trading platforms, there are many problems such as large classification errors, low accuracy in customer interest feature extraction, and high classification time cost. A decision tree algorithm-based customer interest classification method for e-commerce trading platforms is proposed. Based on the basic structure of e-commerce transaction platform web pages, non-target nodes and target nodes are removed, and DFSD fusion method is introduced to extract web browsing content; multi-dimensionally annotate key interest information, and extract customer interest features through multimodal feature fusion; building a customer interest tree for e-commerce trading platforms, pruning customer interest feature data by calculating its entropy value, calculating the information gain of leaf nodes, and building a decision tree classification model based on specific classification rules. Experimental results show that this method reduces classification errors and has good classification results.
  Keywords: decision tree algorithm; e-commerce trading platform; customer interest classification; DFSD fusion method; entropy value; information gain.
  DOI: 10.1504/IJNVO.2024.10059357
   
  
• Driving factors analysis model of social e-commerce platform users' shopping intention based on regression analysis method  
  by Jiangdai Li, Changyi Jin, Jing Zeng 
  Abstract: In order to analyse the driving factors of shopping intention of social e-commerce platform users and improve the effectiveness of user shopping intention analysis, this article proposes a regression analysis based model for analysing the driving factors of shopping intention of social e-commerce platform users. Firstly, the improved synthetic minority oversampling technique (BSMOTE) algorithm is used to sample process the user shopping intention data. Secondly, singular value decomposition (SVD) is used to analyse the data set and shopping intention topics. Then, regression analysis is used to quantitatively describe the driving factors and user shopping intentions, eliminate interfering factors, and construct a model for analysing the driving factors of user shopping intention. The sum of squares decomposition formula is used to modify the model and complete the model construction. Experimental results show that the proposed method can effectively analyse the factors, with an accuracy rate of 95.16% and a recall rate of over 95%.
  Keywords: regression analysis method; BSMOTE algorithm; singular value decomposition; SVD; quantitatively describe; modify the model.
  DOI: 10.1504/IJNVO.2024.10059363
   
  
• A Method for Predicting Consumer Purchase Intention in E-commerce in the Era of Social Media  
  by Aihua Mo 
  Abstract: The research goal is to improve the AUC value and R2 value of the prediction results and reduce the average absolute error, and propose a prediction method of e-commerce consumers' purchase intention in the era of social media. Firstly, linear transformation is performed on the historical purchasing data of consumers, and the importance of the data is determined through thresholds. Based on the judgment results, the data is dimensionally reduced; Secondly, the minimum hash algorithm is used to calculate the similarity between the dimensionality reduced data, and the fuzzy clustering decision method is used to classify historical purchase data; Finally, the Bayesian personalised sorting method is used to predict the purchase intention. The experimental results show that the AUC value and R2 value of the proposed method are large and the average absolute error is low, indicates that the prediction effect of this method is good.
  Keywords: the era of social media; electronic commerce; purchase intention prediction; minimum hash; Bayesian personalised sorting method.
  DOI: 10.1504/IJNVO.2023.10059364
   
  
• E-commerce Collaborative Filtering Recommendation Method Based on Social Network User Relationship  
  by Miao Jiang, Pei Li 
  Abstract: In view of the problems of recommendation methods such as poor recommendation correlation coefficient, high recommendation error, and low strength of social network user relations, research on e-commerce collaborative filtering recommendation method based on social network user relationship is proposed. By analysing data related to user e-commerce platforms, calculating the similarity of social user feature data, and extracting social user feature data. By analysing the factors that affect the trust of social network users, intuitive trust is calculated and multi-dimensional evaluations are conducted to achieve the strength analysis of social network user relationships. Filter e-commerce product data according to e-commerce collaborative filtering model, build e-commerce collaborative filtering recommendation model, and realise e-commerce collaborative filtering recommendation. The test results show that the designed method can improve the correlation coefficient of recommended products, and the recommendation effect is good.
  Keywords: social networks; user relationships; online retailers; collaborative filtering; recommended methods; user characteristics; evaluation matrix.
  DOI: 10.1504/IJNVO.2023.10059365
   
  
• Discussion on the Current Situation and Quality Evaluation Model of Online Course Teaching in Universities Based on Social Network Analysis  
  by WuYi Zhou  
  Abstract: In order to improve the evaluation effect of the teaching quality of modern online courses in universities, we will explore the current situation and quality evaluation mode of online course teaching in universities based on social network analysis. Firstly, analyse social network theory, design a social network analysis process, and then select an online course of Advanced Mathematics from a certain university for research. Use crawler algorithms to obtain interactive data between teachers, students, and students on the platform to achieve community graph visualisation. Finally by analysing the density, distribution, and centrality of community networks, the correlation between centrality and the current situation and quality of curriculum teaching is explored. Based on this result, a quality evaluation model is explored, which can be optimised by introducing social network analysis methods and establishing a scientific student and teacher evaluation mechanism.
  Keywords: social network analysis; SNA; online courses in universities; teaching status; teaching quality evaluation.
  DOI: 10.1504/IJNVO.2023.10059366
   
  
• Analysis of the Impact of Social Media on the Performance of Cross border E-commerce Enterprises from the Perspective of the Digital Economy  
  by Dandan Ye 
  Abstract: In order to explore the impact of social media on the performance of cross-border e-commerce enterprises, this study proposes a study on the impact of social media on the performance of cross-border e-commerce enterprises based on the perspective of digital economy. Firstly, the influencing factors of social media on enterprise performance are selected. Then, the questionnaire was designed. Finally, the long and short term memory neural network algorithm is used to analyse the influence of various factors, and the impact model of social media on the performance of cross-border e-commerce enterprises is built. The experimental results show that social media, entrepreneurial orientation and relationship network all have a positive impact on the performance of cross-border e-commerce enterprises. The more cross-border e-commerce enterprises use social media, the more profits will be increased, which verifies that social media has a direct impact on the performance of cross-border e-commerce enterprises.
  Keywords: digital economy perspective; social media; cross border e-commerce; enterprise performance; long- and short-term memory neural network.
  DOI: 10.1504/IJNVO.2023.10059479
   
  
• Privacy information encryption for cross-border e-commerce users based on social network analysis  
  by Na Wang, Feng Gao, Ji Zhang 
  Abstract: In order to protect the privacy information of cross-border e-commerce users, an encryption algorithm based on social network analysis is proposed in this paper. Firstly, the logical inference mapping method for blockchain identity data is used to encode public and private keys, and the asymmetric encryption method is applied to construct keys. Then, the social network analysis method is used to rearrange the user social network structure, and the user information fusion processing and optimised encryption are realised with the support of arithmetic coding, homomorphic encryption and other technologies. The simulation results show that this method has strong anti-attack ability and low time cost.
  Keywords: social network analysis; cross-border e-commerce; privacy; information encryption; vector quantisation encoding.
  DOI: 10.1504/IJNVO.2023.10059480
   
  
• Dynamic Grouping Method for Online Learning Behavior Based on Social Network Analysis  
  by Juan Guo 
  Abstract: In this paper, a dynamic grouping method for online learning behaviour based on social network analysis is proposed. The Louvain community detection algorithm is utilised for social network analysis. Based on the analysis results and distributed web crawler, online learning behaviour information is mined and subjected to standardised processing, including outlier removal, noise filtering, and data segment alignment, to extract relevant features. A grouping objective function based on the XGBoost algorithm is constructed for the dynamic grouping of online learning behaviour. The objective function is solved to obtain the dynamic grouping results. Experimental results demonstrate that the proposed method achieves a minimum relative error rate of 1.3% in feature extraction, a maximum accuracy of 97.9% in grouping, and an average task completion time of 0.77 s.
  Keywords: social network analysis; online learning behaviour; dynamic grouping; distributed web crawler; XGBoost algorithm; objective function.
  DOI: 10.1504/IJNVO.2024.10060291
   
  
• Evaluation of Teaching Effectiveness in Higher Education Based on Social Networks  
  by Xiyang Li, Quanzhong Yang 
  Abstract: The evaluation of educational and teaching effectiveness is beneficial for universities to understand the current teaching situation, formulate reasonable teaching strategies, and improve teaching quality. Therefore, a social network-based evaluation method for educational and teaching effectiveness in universities is proposed. Firstly, analyse the role of teaching effectiveness evaluation and design the principles for constructing a teaching effectiveness evaluation index system; then, calculate the entropy matching degree of evaluation indicators, set consistency criteria for evaluation indicators, and use SVM algorithm to classify evaluation indicators, constructing an evaluation indicator system; Finally, construct a social network for teaching effectiveness evaluation, determine the subgroup cohesion of evaluation indicators, construct an education and teaching effectiveness evaluation function, and obtain the evaluation results. The experimental results show that the confidence level of the evaluation results of this method is high, and the matching error of the entropy value of the evaluation indicators is low.
  Keywords: social network; the effectiveness of higher education and teaching; indicator system; entropy matching degree.
  DOI: 10.1504/IJNVO.2024.10060321
   
  
• Research on online free marketing mode based on social network analysis  
  by Chang Liu, Bo An 
  Abstract: In order to control the cost of online free marketing mode, a recommendation model for online free marketing mode based on social network analysis is proposed in this paper. Firstly, a set of cost constraint indicators for online free marketing mode is constructed, and the homomorphic parameters are processed by multi-dimensional parameter estimation fusion. Then principal component analysis and social network analysis are conducted for block constraints to the dynamic cost characteristics, and the cost correlation analysis of the online free marketing mode is also conducted to realise the optimal cost control solution. The simulation results show that the proposed method can reduce the cost, maximise the data traversal and minimise the computational complexity.
  Keywords: social network analysis; online free marketing mode; supply chain; comprehensive coordinated control; homomorphic fusion; principal component analysis.
  DOI: 10.1504/IJNVO.2024.10060322
   
  
• An Improved Data Aggregation for Fog Computing Devices in Internet of Things  
  by M. Jalasri, L. Lakshmanan 
  Abstract: Diverse data that have private information are stored in the cloud by institutions and users. Fog computing has progressed with regard to service latency and has also been comprehensively studied. A proposal has been made in this work to secure IoT data, a clustering algorithm that utilises backup cluster head (CH) for network performance enhancement, and also utilises a multi-route protocol for data transmission towards the fog system. Comparisons are made between the proposed algorithm and the energy efficient heterogeneous clustering algorithm (EEHCA). A novel particle swarm optimisation (PSO) and river formation dynamics (RFDs) algorithm was proposed in this work for effective CH election in wireless sensor network (WSN) to efficiently transmit data towards the base station (BS) with minimised energy. This technique bypasses local optima during the solution search, leading to significantly improved results. RFD-EEHCA outperforms EEHCA by 8.53% and PSO by 8.09% in terms.
  Keywords: cloud computing; internet of things; IoT; fog computing; security; wireless sensor network; WSN; routing; clustering.
  DOI: 10.1504/IJNVO.2024.10060435
   
  
• Topic text detection by clustering algorithm for social network media  
  by Sha Sha 
  Abstract: The advent of the Internet era has promoted the development of social network media, making the number of people active in these social network platforms greatly increase, and the resulting large amount of data and information makes the fast location retrieval of topics of interest a problem. This paper detected topic texts of social network media by the modified particle swarm optimisation-based K-means (MPSO-means) clustering algorithm to achieve topic text clustering effect and alleviate the problem of inconvenience caused by information overload. The results of the study showed that the clustering results of short texts showed a trend of outperforming long texts; and the MPSO-means algorithm was closer to 1 than the other two algorithms in the values of silhouette coefficient, clustering purity, and homogeneity, with better clustering effect, and also consumed the shortest time in detection, only 1,196 s.
  Keywords: text clustering; social network media; topic text; modified particle swarm optimisation-based K-means.
  DOI: 10.1504/IJNVO.2024.10060582
   
  
• A Knowledge Set Recommendation Method for Online Education in Universities Based on DV-TransE Model and Social Networks  
  by Die Meng, Beibei Ma, Zhanlei Shang 
  Abstract: In order to improve the recommendation accuracy of existing online education knowledge sets in universities and shorten the recommendation response time, a recommendation method for online education knowledge sets in universities based on DV-TransE model and social network is proposed. This method is first based on the principle of knowledge graph, extracting descriptive features of the knowledge set, and introducing the TransE algorithm to construct the DV-TransE model of the online education knowledge set in universities. Then, based on social networks, the similarity between users is calculated, and finally, it is combined with the constructed knowledge set DV-TransE model to achieve recommendation of online education knowledge sets in universities. The experimental results show that after the application of the proposed method, its recommended response time is less than 14.5 ms, and the recommendation accuracy is as high as 95%, which is superior to the comparison method.
  Keywords: DV TransE model; social networks; online education; knowledge set; recommended methods.
  DOI: 10.1504/IJNVO.2024.10060713