Skip to Main Content
 

Global Search Box

 
 
 
 

Files

ETD Abstract Container

Abstract Header

Latent Factor Models for Recommender Systems and Market Segmentation Through Clustering

Abstract Details

2017, Master of Science, Ohio State University, Statistics.
Consider the problem of recommending products to a set of online users, where a very large selection of potential products are available. Recommender systems were introduced to predict users preferences and to provide users with personalized recom- mendations on products or services. These systems use collaborative ltering methods to analyze user preferences based on past behaviors, for instance, their ratings. In the collaborative ltering framework, most of the ratings are missing and the goal is to recover all of the unobserved ratings. This thesis investigates the performance of the most widely-used class of models in collaborative ltering, known as latent factor models, in performing collaborative ltering on a dataset set of Amazon Fine Food Reviews. Latent factor models factorize user ratings of products into user and item feature vectors. By taking this approach, the models aim to nd low-rank feature matrices to describe the data. This is accomplished by a penalized least squares loss function which is minimized using the Stochastic Gradient Descent searching technique. A potential approach to extract meaning from these feature matrices is to apply clustering techniques. A heuristic clustering method is proposed to cluster the items based on their hidden characteristics that are identi ed by the models. To further ex- amine the similarities within each cluster and dissimilarities among di erent clusters, text mining is employed on the corresponding text reviews.
Matthew Pratola (Advisor)
Laura Kubatko (Advisor)
Radu Herbei (Committee Member)
77 p.

Recommended Citations

Citations

  • Zeng, J. (2017). Latent Factor Models for Recommender Systems and Market Segmentation Through Clustering [Master's thesis, Ohio State University]. OhioLINK Electronic Theses and Dissertations Center. http://rave.ohiolink.edu/etdc/view?acc_num=osu1491255524283942

    APA Style (7th edition)

  • Zeng, Jingying. Latent Factor Models for Recommender Systems and Market Segmentation Through Clustering. 2017. Ohio State University, Master's thesis. OhioLINK Electronic Theses and Dissertations Center, http://rave.ohiolink.edu/etdc/view?acc_num=osu1491255524283942.

    MLA Style (8th edition)

  • Zeng, Jingying. "Latent Factor Models for Recommender Systems and Market Segmentation Through Clustering." Master's thesis, Ohio State University, 2017. http://rave.ohiolink.edu/etdc/view?acc_num=osu1491255524283942

    Chicago Manual of Style (17th edition)