Arithmetic coding based string approximation
Mayank Sharma · 2017
Arithmetic coding gives lowest range of entropy for any string computed form joint probability value of sub-strings of keyword. It is very much permissible to use this technique for mapping a large set of keywords to real values with in the range of [0; 1]. This will create a space of search containing tag value range for keywords and mapped all the keywords to this space. In this paper a new string approximation technique is proposed, which is based on the arithmetic coding used for mapping the misspelled keyword to the tag value interval in a limited search space containing the nearest possible keywords against the misspelled keyword. Using such technique time for searching all the nearest keywords is dramatically reduced due to reduction of search space.