Biomedical Text Summarization Based on the Itemset Mining Approach

2. Background

Various methodologies with machine intelligence, linguistics, and statistics are projected to create biomedical document summarization frameworks (Gambhir & Gupta, 2017; Yao et al., 2017). Numerous summarization systems evaluate significance with strings dependent on the top of conventional estimates, for example, sentence length, term position, sentence location, cue phrase, etc. (Gupta & Lehal, 2010). In this biomedical text summarization, various performances have revealed that such general features can perform better for the domain-specific methods (Moradi, 2018b; Moradi & Ghadiri, 2018; Plaza et al., 2011). The biomedical corpus has its uniqueness (Erhardt et al., 2014; Plaza et al., 2011) which can cause some trouble using area-independent summarization techniques. This acted as the main impetus to create domain-specific summarizers which can counter the distinctive biomedical document. In this research, biomedical text summarization utilizes a preprocessing step where input text is designed, including the UMLS metathesarus (Nelson et al., 2017). Many different techniques are uses for comparative study on opinion summarization in extractive and abstractive way with the help of dimensionality reduction (Bhatia S, 2020).