In every moment, there is a huge capacity of data and information communicated through social network. Analyzing huge amounts of text data is very tedious, time consuming, expensive and manual sorting leads to mistakes and inconsistency. Document dispensation phase is still not accomplished of extracting data as a human reader. Furthermore the significance of content in the text may also differ from one reader to another. The proposed Multiple Spider Hunting Algorithm has been used to diminish the time complexity in compare with single spider move with multiple spiders. The construction of spider is dynamic depends on the volume of a corpus. In some case tokens may related to more than one topic and there is a need to detect Topic on semantic way. Multiple Semantic Spider Hunting Algorithm is proposed based on the semantics among terms and association can be drawn between words using semantic lexicons. Topic or lists of opinions are generated from the knowledge graph. News articles are gathered from five dissimilar topics such as sports, business, education, tourism and media. Usefulness of the proposed algorithms have been calculated based on the factors precision, recall, f-measure, accuracy, true positive, false positive and topic detection percentage. Multiple Semantic Spider Hunting Algorithm produced good result. Topic detection percentage of Spider Hunting Algorithm has been compared to other algorithms Naïve bayes, Neural Network, Decision tree and Particle Swarm Optimization. Spider Hunting Algorithm produced more than 90% precise detection of topic and subtopic.
Topic Detection Using Multiple Semantic Spider Hunting Algorithm
