Dual Access Cache Memory Management Recommendation Model Based on User Reviews

With the growing number of Internet services, the prediction of relevant services based on the user opinions is the major task in the recommendation model. The inclusion of important information taken from the user-generated textual reviews (in textual comments) is the major objective in the diverse review-based recommendation. An advanced textual analysis extracts the different elements for recommendation such as reviewer review, contextual information and the comparative opinions to improve the content-based recommendation performance. The major problems in such content-based recommendation are sparsity and memory management. Hence, this paper proposes the Dual Accessing Cache Memory Management (DACMM) to alleviate the issues in the recommendation system. The provision of ratings to the products based on the user experiences and their learning contribute to recommend the interesting product to the end users. With an increase in the dimensionality of products, the dimensionality of their reviews is increased. This paper employs the feature extraction that utilizes the positive, negative reviews with the overall word count to identify the features. The assigning of the index corresponds to the user review with the positive and negative features reduce the size of products into the interesting category. The integration of query processing with the cache memory management reduces the complexity in recommendation operation effectively. The comparative analysis between the proposed DACMM with the existing clustering and cache management models in terms of query latency, hit ratio, Mean Absolute Error (MAE), computational time and precision assure the effectiveness of the proposed DACMM in the review-based recommendation.