One-class Collaborative Filtering with Temporal Information

Singh, Shashank (2017) One-class Collaborative Filtering with Temporal Information. Masters thesis, Indian Insitute of Technology Hyderabad.

Text
CS15MTECH11013.pdf - Submitted Version
Restricted to Registered users only until 16 July 2020.
Download (1MB) | Request a copy

Abstract

In One-Class Collaborative Filtering (OCCF) only positive examples or implicit feedback can be observed. The one-class feedback in the form of (user, item) pairs is easily available than numerical ratings in the form of (user, item, rating) triplets as exploited by traditional collaborative filtering algorithms. The training data usually consists of binary data, reflecting a users action or inaction, such as buying items from the store, on E-commerce websites or web-page bookmarking in the book- marking scenario. The dataset in systems like above have very few entry. The user-item pair for which entry is there, we treat it as a positive example. Negative examples and unlabeled positive examples are mixed and we are typically unable to distinguish them. In most of systems providing one-class feedback, temporal information about the items, users are also available. In my Thesis, I have considered the temporal information related to items(release date) and users(item purchase time) and popularity of items, to add negative examples for users and study the performances of di ↵ erent new time-aware algorithms on several benchmark datasets. KEYWORDS: Collaborative Filtering, Stochastic Gradient Descent, One Class Collaborative Fil- tering, Non-negative Matrix Factorization, Weighted alternating least Square, Temporal Dynamics.

[error in script]

IITH Creators:

IITH Creators	ORCiD

Item Type:

Thesis (Masters)

Uncontrolled Keywords:

collaborative filtering, stochastic gradient descent, one-class collaborative fildering, non-negative matrin factorization, Temporal dynamics, weighterd alternating least square, TD911

Subjects:

Computer science

Divisions:

Department of Computer Science & Engineering

Depositing User:

Team Library

Date Deposited:

17 Jul 2017 05:44

Last Modified:

17 Jul 2017 05:44

URI:

http://raiithold.iith.ac.in/id/eprint/3389