~ Semantic Search Art ~

Home Downloads Services Research Resources Contact Semantic search is the next step after keyword search. It is part of natural language processing (NLP) and sometimes is called mathematical linguistics. NLP itself is a very wide area, it may include such non-traditional computer applications as detecting deception in speech, but semantic search is only part of it and is a fast way of detecting text fragments conveying the same information but not necessarily with the same words. The field of application is web search and desktop search. In these cases the text fragments are technical, business or fiscal documents 2 to 20 pages, when printed, and dedicated to a single and rarely to several topics that may not be clearly declared in text and on that reason called latent. This site provides review and explanation of several technologies Latent Semantic Analysis Probabilistic Latent Semantic Analysis Hierarchical Agglomerative Clustering Naive Bayes Support Vector Machine Latent Dirichlet Allocation k-means clustering k-nearest neighbors n-grams Random Forest Self-Organizing Maps Inverted File Index Conceptual Search Polar Orthogonalization Comparison of the best clustering technologies Feature Extraction Split and Merge Predictive Coding plus data corpus and software allowing to reproduce experiments. The links above point to the test projects published on this site under menu item RESEARCH. RESOURCES provide list of semantic search engines and companies conducting research, development and sale of software and list of recent and most interesting patents granted in United States. Section DOWNLOADS contains utilities for semantic search. Some of them are freely available and some of them available for the charge.