OPERATION ALGORITHM OF SELF-TAUGHT HASHING BY SEARCH OF SIMILAR DOCUMENTS

Quick search by similarity among big amount of data is a current issue for many modern applications, especially for web search. Semantical hashing is the most optimal way for it. One of the new methods based on it is called Self-Taught Hashing by proposition of Chinese scientists. This method has shown better results in speed and quality than analog methods, and has its own peculiarities of algorithm work as well.