Abstract
Abstract With the advent of the era of big data, our lives generate huge amounts of data every day, and the field of e-commerce is no exception. It is particularly important to analyze these data and recommend products. It is reported that through the recommendation algorithm, Amazon has increased its sales by about 30%. Among the recommended algorithms, the collaborative filtering algorithm is currently relatively mature and has achieved very good results in various fields. But the traditional collaborative filtering algorithm is too rough when calculating the similarity and prediction score, and the efficiency is very low. We combine the traditional collaborative filtering algorithm with the decision tree algorithm, and improve the traditional recommendation algorithm, create a collaborative filtering decision tree algorithm to recommend products, and run the new collaborative filtering decision tree algorithm on the Hadoop platform on. Experiments show that the improved algorithm makes the accuracy of recommendation significantly improved.
Highlights
With the development of science and technology, Internet technology has rapidly developed and popularized, so that the data on the network is growing at the level of PB every day, bringing a lot of information resources to users and greatly enriching people's daily lives
The recommendation algorithm combines the decision tree and the collaborative filtering algorithm, and improves the traditional collaborative filtering algorithm to improve the timeliness of recommendation
There are two parameters that can be controlled by the random forest model, one is the number of random forest decision trees, and the other is the number of feature attributes that are randomly extracted to build the decision tree
Summary
With the development of science and technology, Internet technology has rapidly developed and popularized, so that the data on the network is growing at the level of PB every day, bringing a lot of information resources to users and greatly enriching people's daily lives. The recommendation system is automatic and intelligent to recommend items for users, and it will dynamically adjust the recommended item types according to the changes of user behavior, which truly avoids the "information overload" problem. Faced with such a huge amount of data, it is necessary to adopt a big data model for analysis. MapReduce is a distributed computing framework under Hadoop [2] It uses the "divide and conquer" idea to decompose complex tasks or data into several simple tasks for parallel processing. The recommendation algorithm combines the decision tree and the collaborative filtering algorithm, and improves the traditional collaborative filtering algorithm to improve the timeliness of recommendation
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
