Abstract:Query suggestion of Web search is an effective approach to help users quickly express their information need and accurately get the information they need. Most of popular web-search engines provide possible query suggestions based on their query log data, which is a kind of implicit relevance based approach. However, it is difficult to give suggestions to search queries that have no or few historical evidences in query logs. To solve this problem, traditional pseudo relevance based approaches directly extract additional keywords from the top-listed search results of a given search query as suggestions. However, for hot topic or event related search queries, users more like to browse the latest and newly appeared contents. In this paper, we follow the direction of pseudo relevance based suggestion approaches by mining microblog data that is inherent in fast information propagation and dissemination. Our graph based rank aggregation approach combines a frequency based ranking with considering words themself and a LDA (Latent Dirichlet Allocation) based ranking by mining hidden topics behinds words. A dataset is crawled from the posts of fourteen micro-topics of Sina microblog platform. The experimental results clearly demonstrate our proposed approach is more effective than traditional pseudo relevance based methods. Moreover, the suggested keywords extracted from the posts published by authenticated users are more effective than two traditional pseudo relevance based approaches, i.e., the posts submitted by all users and the top returned posts returned by Sina search engine. In addition, applying LDA on microbog posts alone is far from satisfactory, but the combination of the frequency based ranking and the LDA based ranking show much better performance.

Real-Time Search over a Microblogging System

Real-time and Personalized Search over a Microblogging System.

A Real-Time Distributed Index Based on Topic for Microblogging System.

Socialanalysis: A Real-Time Query And Mining System From Social Media Data Streams

LiveIndex: A Distributed Online Index System for Temporal Microblog Data

Microblog Search and Filtering with Real-Time Dynamics Based on BM25

Compact Indexing and Judicious Searching for Billion-Scale Microblog Retrieval.

Real-time Recommendation for Microblogs

Microblog Search and Filtering with Time Sensitive Feedback and Thresholding Bsed on BM25.

Learning to Rank Microblog Posts for Real-Time Ad-Hoc Search.

A System to Manage and Mine Microblogging Data.

Enhancing Keyword Suggestion of Web Search by Leveraging Microblog Data.

The Design and Implementation of Micro-Blog User Interest Search Engine Base on Cloud Computing Technology

Real-Time and Personalized Recommendation on Microblogging Systems

Real-time Personalized Twitter Search Based on Semantic Expansion and Quality Model

Towards Effective Event Detection, Tracking And Summarization On Microblog Data

Microblog Sentiment Orientation Detection Using User Interactive Relationship

Top-k Temporal Keyword Search over Social Media Data

On Improving a Microblog Ranking.

A Temporal Microblog Filtering Model

A Time-Sensitive Model for Microblog Retrieval