FUZZY SCORE BASED SHORT TEXT UNDERSTANDING FROM CORPUS DATA USING SEMANTIC DISCOVERY

Short text understanding and short text are always more ambiguous. These short texts are produced including Search queries, Tags, Keywords, Conversation or Social posts and containing limited context. Generally short texts do not contain sufficient collection of data to support many state-of-the-art approaches for text mining such as topic modelling. It presents a comprehensive overview of short text understanding. Here we used a novel framework are Text Feature Extraction Algorithm and Fuzzy weighted Vote algorithm First, Text classification based on semantic feature extraction. Its goal is that use semantic feature extraction to improve the performance of classifier. And second, Fuzzy weighted Vote algorithm is the combination of Fuzzy logic and weighted vote algorithm, which means it generates the fuzzy score and then based on this score the weight is calculated during shortening the text. In experimental results, the novel Feature Extraction and voter has higher safety performance than the previous classification algorithms. This proposed criterion can provide almost accurate safety and also a good range of accessibility. We have proved that in problems where the weighted voting distinguish some alternatives and finds the best alternative. Reduced Computation time comparing to other previous process and schemes