In the era of big data where there are superfluous data in the web pages to be surfed, it become challenging for e the service provider of the ecommerce to identify the user need in efficient manner. Also, these days the user want to have products based upon their behavior. Hence, this paper presents a recommender system together with the machine learning for analyzing the behavior of the user over web page traversing and providing them most adequate web page having the product of his or her choice. The paper also tries to include most relevant algorithms that can drive user analysis more appropriately over the server logs that has bene triggered by the user. The paper uses the Singular Value Decomposition (SVD) to understand the weightage of the user-product interaction in order to provide more accurate results for the web-usage analysis for recommending most suitable product. The paper also takes into account the redundancy of data in user-product interaction that may have occurred due to technical reason, which make this work more efficient when recommending product.