Word Embeddings for User Profiling in Online Social Networks

Abstract

User profiling in social networks can be significantly augmented by using available full-text items such as posts or statuses and ratings (in the form of likes) that users give them. In this work, we apply modern natural language processing techniques based on word embeddings to several problems related to user profiling in social networks. First, we present an approach to create user profiles that measure a user's interest in various topics mined from the full texts of the items. As a result, we get a user profile that can be used, e.g., for cold start recommendations for items, targeted advertisement, and other purposes; our experiments show that the interests mining method performs on a level comparable with collaborative algorithms while at the same time being a cold start approach, i.e., it does not use the likes of an item being recommended. Second, we study the problem of predicting a user's demographic attributes such as age and gender based on his or her full-text items. We evaluate the efficiency of various age prediction algorithms based on word2vec word embeddings and conduct an extensive experimental evaluation, comparing these algorithms with each other and with classical baseline approaches.

Building Semantic User Profile for Polish Web News Portal

“…As found in [3], for predicting demographic data with Word2Vec-based user profiles, models trained with a custom text dataset yield better results than those using a large general language corpus. Considering also the dynamic nature of news content and changing user interests, we verify this claim and use a Word2Vec model trained on a corpora consisting of Polish Wikipedia and National Corpus of Polish Language NKJP [25] to compare with our models built on a much smaller custom corpus with the same model parameters.…”

“…Among the frameworks applied to demography prediction, there are both topic modeling approaches and Word2Vec representations. Moreover, in [2,3], the authors use Word2Vec-based profiles from social media texts for predicting a user's age and recommendations. We also apply profiles based on Word2Vec word representations and compare their results with the topic modeling approach for predicting user gender.…”

“…Considering also the dynamic nature of news content and changing user interests, we verify this claim and use a Word2Vec model trained on a corpora consisting of Polish Wikipedia and National Corpus of Polish Language NKJP [25] to compare with our models built on a much smaller custom corpus with the same model parameters. As the text preprocessing tools used by [3] are different from ours, we compare models trained on all word forms. The corpora descriptions are presented in Table 1.…”

“…Thus, the resulting user profile from LDA representations describes the average user's interest in each of the latent topics, while the Word2Vec profile is constructed analogically to the article representation in Section 2.3.2 -assuming the linear characteristics of the skip-gram model [23], the average article vector also approximates the user's semantic interests. A similar approach was proposed for age prediction with Word2Vec-based profiles for social networks in [3].…”

Meta-User2Vec model for addressing the user and item cold-start problem in recommender systems

“…Moreover, there have been a few attempts to apply text embedding methods to represent content-based user profiles. In Musto et al (2016); Alekseev et al (2017); Misztal-Radecka (2018), Word2Vec is used to build content-based user profiles. It was observed that this approach gives comparable results to the standard collaborative filtering techniques, especially for sparse datasets.…”