Research on Recommendation Algorithm Based on Ranking Learning

Research on Recommendation Algorithm Based on Ranking Learning

Xiaoli Zhang (Hetao College, Bayannur, China)
Copyright: © 2019 |Pages: 14
DOI: 10.4018/JECO.2019010106


After analyzing the logistic regression and support vector machine's limitation, the author has chosen the learning to rank method to solve the problem of news recommendations. The article proposes two news recommendation methods which were based on Bayesian optimization criterion and RankSVM. In addition, the article also proposes two methods to solve the dynamic change of user interest and recommendation novelty and diversity. The experimental results show that the two methods can get ideal results, and the overall performance of the method based on Bayesian optimization criterion is better than that based on RankSVM.
Article Preview

1. Introduction

With the development of Internet and information technology, the people get information way has changed greatly. For example, people used to rely on traditional news media (such as newspapers, radio, television, etc.) to obtain news information, and later transformed into the main use of Internet media. Later, used Internet media is to obtain news information (Huang, Zhang, Tian, Sun, Xiangyang, 2016; Meng, Chen, Zhang, 2016; Ding, 2013). This change makes it easier for people to get information. In recent years, the emergence and rapid development of social media have further changed people's live, entertainment and other ways.

The emergence of social media is to people's lifestyle impact mainly in two aspects:

  • 1.

    The people get information way has changed dramatically. Before the advent of social media, people are passive information recipient. At this time people mainly by reading newspapers, listening to radio and television, browsing the web, search related content and other ways to passively accept news. Rarely participate in the creation, release and dissemination of information. With the advent of social media as the representative of the Web2.0 era, people are also from the passive recipients of information into information publishers, communicators and recipients. This change allows each user to release and disseminate information through the network media, so that each ordinary user can become a “journalist”. So that, each user can easily and quickly share their views and information at any time.

  • 2.

    People's way of communication has undergone major changes

Twitter, Facebook, Sina, and micro-blogs represented by social media provide users with a platform. Users can publish their own state information, pay attention to their interested users can timely access to the relevant information and by forwarding and comment to the sharing of information and interaction with friends using the platform. At the same time, the social platform also narrowed the distance between people and the world had no intersection of two users through the platform provided by the attention, to establish contact to add friends and other functions, and then based on the links for information is sharing, interactive discussion (Xiao, 2014; Han, 2014).

Since it’s not possible to obtain the explicit feedback information about whether user loves one piece of news (Wang, 2013; Fang, 2011; Lian, 2014; Cai, 2011; Wang, 2015), so it needs to build user’s feedback dataset in one manner. The optimal way is to acquire accurate dataset by means of user’s rating or the level of rating about historical data (Song, 2015; Wang, 2015; Zhou, 2015; Tang, 2015). But the way costs too much and is not applicable in practice. To break through the constraint, the paper attempts to use sequential learning method to solve the question of news recommendation (Xie, 2014; Meng, Hu, Wang, Zhang, 2013; Li, Wang, He, Jin, 2015; Li, 2014; Peng, Sun, Han, Chen, 2015) of user’s interest or no in one piece of information to the one of judging the size of user’s interest in two pieces of news; next, based on the dataset, employ sequential learning method to train the model as further to predict user’s interest (Lu, 2013; Cao, 2013) Besides, the paper explores into the question of user interest dynamism and recommendation diversity and novelty faced in the field of recommendation (Wen, Cai, Wu, 2014; Ma, 2016)

In order to solve the problem of user information overload, this paper uses the method of personalized ranking of information flow to solve. Social media is usually the flow of information in accordance with the time sequence of information to sort. This sort of way for the user to present the latest friend status information, but there is no combination of user interest. This approach will adversely affect the user experience.

The purpose is: the user is interested in micro-blog ranked in the location of the user information flow, the user is not too interested in micro-blog ranked in the opposite position.

Complete Article List

Search this Journal:
Open Access Articles
Volume 18: 4 Issues (2020): 1 Released, 3 Forthcoming
Volume 17: 4 Issues (2019)
Volume 16: 4 Issues (2018)
Volume 15: 4 Issues (2017)
Volume 14: 4 Issues (2016)
Volume 13: 4 Issues (2015)
Volume 12: 4 Issues (2014)
Volume 11: 4 Issues (2013)
Volume 10: 4 Issues (2012)
Volume 9: 4 Issues (2011)
Volume 8: 4 Issues (2010)
Volume 7: 4 Issues (2009)
Volume 6: 4 Issues (2008)
Volume 5: 4 Issues (2007)
Volume 4: 4 Issues (2006)
Volume 3: 4 Issues (2005)
Volume 2: 4 Issues (2004)
Volume 1: 4 Issues (2003)
View Complete Journal Contents Listing