Web Information Retrieval Lab @ Taipei Tech

People

Faculty

Jenq-Haur Wang

Research Interests

Big Data Analytics and Deep Learning in Different Domains

In previous MOST projects, we constructed a distributed platform for big data analytics. In this platform, we continue to develop applications in various domains such as social network analysis, crime analysis, and smart manufacturing in Industrie 4.0. We plan to conduct statistical and content analysis in different domains. Major issues are addressed and various machine learning and deep learning methods are compared and evaluated

Web Information Mining and Recommendation for Smart TV

In our previous three-year NSC project, we investigated the key technologies for Hybrid Internet Broadcast TV (HIBTV).

We focus on integrating and analyzing information from both TV and the Internet, and giving useful recommendations for users.

In the next three years, we are going to improve relevant concept and opinion mining using big data analytics platform for giving context-aware recommendation on smart TV.

Opinion Mining and Sentiment Analysis in Social Media

During the past few years, we have been analyzing social media content in microblogs for sentiment analysis and opinion mining.

Palanteer (TW Election) is a social search system that analyzes plurk data from a set of Taiwan users.

Targeting at the 2012 Presidential Election in Taiwan, this system was created in collaboration with LARC (Living Analytics Research Centre), Singapore Management University.

We have been collaborating with LARC, SMU in many real-life applications with social media analytics.

By exploiting both social media content and structural information such as replying, forwarding, and likes, we are able to identify the most influential and popular topics and extract user opinions.

Currently, we are applying related technologies in various applications such as recommender systems for drama, movies, and tourism on smart TVs.

Large-Scale Near-Duplicate Copy Detection and Tracking for Multimedia Content

Integrating information retrieval techniques with networking technologies, we proposed a distributed indexing approach to large scale multimedia content archiving, retrieval, near-duplicate detection and tracking.

Based on this framework, we investigated the effect of multimedia content protection for texts, images, and videos.

In future, we are planning to address the issues in multimedia content tracking for online videos.