User-Guided Aspect Classification for Domain-Specific Texts

• DOI: 10.48550/arxiv.2004.14555
• Publication Date: 2020-01-01
• Authors: Peiran Li; Fang Guo; Jingbo Shang
• Keywords: FOS: Computer and information sciences; Computer Science - Computation and Language; 0202 electrical engineering, electronic engineering, information engineering; 02 engineering and technology; Computation and Language (cs.CL)

Aspect classification, identifying aspects of text segments, facilitates numerous applications, such as sentiment analysis and review summarization. To alleviate the human effort on annotating massive texts, in this paper, we study problem classifying based only a few user-provided seed words for pre-defined aspects. The major challenge lies how to handle noisy misc aspect, which is designed texts without any Even domain experts have difficulties nominate making existing seed-driven classification methods not applicable. We propose novel framework, ARYA, enables mutual enhancements between aspect via iterative classifier training updating. Specifically, it trains then leverages induce supervision aspect. prediction results are later utilized filter out Experiments two domains demonstrate superior performance our proposed well necessity importance properly modeling