Header menu link for other important links
Survey of Classical and Advanced Techniques for Product Recommendations
Published in JETIR
Volume: 6.0
Issue: 5.0
Pages: 531 - 534

Extensive use of data and increasing sales by implementing recommender system by various retail and e-commerce companies through their website is tremendously boosted in current scenario. Recommendation systems is a software application that provide suggestions to the intended user regarding which products is buy and which is not to buy. This systems uses various filtering techniques to provide recommendations to the customers. Filtering techniques that helps in decision making of customers are most likely content-based filtering, CNN model filtering and hybrid filtering. In order to reduce information search time of user it is need to integrate the recommended facility in E-Commerce. Content-based filtering is used due to its suitability in the domain or in situations where the products are more than users. IDF Weighted word to vector model were used to determine whether a product is relevant or similar to a user's profile of interest. CNN model is used to find out image based similarities and both of this technologies integrated as a hybrid system and finally system give recommendation based on hybrid filtering system. This paper also gives an importance of an algorithm for providing recommendations or suggestions based on queries of users. Algorithms employ both IDF Weighted Word to vector model and CNN model. This review gives an overview of available data sets, methods for preprocessing data sets, recommendations techniques and challenges involved.

About the journal
JournalJournal of Emerging Technologies and Innovative Research
Open Access0