B-Tech Computer Science Final Year Project
We propose to solve the link prediction problem in sparse matrix using clustering with association rule mining. The model learns latent features from the sparse matrix, and makes better predictions. To present the effect of clustering the data onto the association rules. Hence, we have compared the results of two different approaches: Finding association rules without consumer segmentation, and with consumer segmentation. The data analysis framework is applied to the data of Online Retail. By extracting the most important information from Online Retail data, we claim that this framework provides offers, the right product/advertisement to the right consumer. Predicting from a sparse matrix using algorithms like ARM are quite expensive and it takes a lot of time and space but using Clustering with ARM we can predict in Big O(n)+2k time where k is frequent item sets.
Project(1).pdf contains all the information and ways in which we carried out our project in detailed manner.
Final_Year_Project.pdf is our final paper covering all details in concise manner.