High-Throughput Linguistic Content Sentiment Analysis

Explosive growth of social media has transformed how people communicate, interact, and actively express their opinions about different topics. Scrawlr’s unique model for platform management allows extensive freedom for users to generate their content, creating a novel opportunity to evaluate user opinions and network structure. A popular method to analyze online content is sentiment analysis. While research on sentiment analysis is growing explosively, most methods rely on lexicon-based or machine learning approaches. The majority of research efforts are designed to work with only English content, while a significant share of information is available in other languages. In the proposed research, using machine learning algorithms, we develop an automated content sentiment analysis in multiple languages and take a different step into this field, which is providing the capacity to enable comparison of sentiment conclusions against the evaluation and classification of content by users. In other words, we train a set of data, then predict the labels for the rest based on the train set.

Faculty Supervisor:

Ketra Schmitt

Student:

Partner:

Scrawlr Development Inc.

Discipline:

Computer science

Sector:

Information and Communications Technology; Artificial Intelligence; New and Digital Media

University:

Concordia University

Program: