An Efficient Converse Mapping Technique in Modern Information Retrieval
M Ramya, E. Thenmozhi · 2014
A traditional forward dictionary maps from words to their definitions. Unlike a regular dictionary, the reverse dictionary performs converse mapping that maps from definitions to words. The user input phrase describes the definition of desired concept, and returns an efficient candidate word that satisfies the input phrase. The reverse dictionary addresses the widespread problem of knowing the meaning of a word, but unable to recall the appropriate word on demand. The converse mapping technique finds the exact result for the user entered keyword by comparing the partitioned input with dictionary database. The partitioning method increases the overall scalability and distributes the data across multiple threads. The efficiency of the reverse dictionary can be improved by reducing the set of definitions in the comparison process. The query expansion technique is used to improve the potential of reverse dictionary and increases the probability of identifying relevant definition. The approaches of reverse mapping provide significant improvements in the performance scale. The converse mapping technique in modern information retrieval extracts the best matched result without sacrificing the quality of the solution. Latent Semantic Indexing is a dimensionality reduction technique which reduces the descriptive length of the document statistical structure. Principal Component Analysis is a statistical procedure that converts the correlated variables into a set of values of linearly uncorrelated variables. Latent Dirichlet Allocation (LDA) and Probabilistic Latent Semantic Indexing (PLSI) are used for building efficient schemes in reverse dictionary. LDA finds the short description from a large collection of document while preserving the essential statistical relationship. The approach of LDA reduces the descriptive length of document that in turn provides modularity and extensibility. PLSI defines a generative model of the data and minimizes word perplexity. The standard techniques from statistics can be applied for combining the models and controlling the complexities. The reverse mapping set maps for all terms appearing in the sense phrase and determine the most generic form by reducing various inflexions of the same word to a common form. The converse mapping technique uses stepwise refinement approach for extracting the best matched result for an input phrase. The approach on information retrieval finds the core terms from the sense phrase and searches for the candidate words whose definitions containing the similar core terms. The synonyms, hyponyms and hypernyms of the terms increase the probability of obtaining sufficient number of outputs. The results are sorted by comparing the input with every definition in the dictionary database. The similarity measure between the sentences determines the importance of terms that contribute more to the meaning of a phrase. In modern information retrieval, the converse mapping technique can provide significant improvements in performance and extracts efficient result without impacting available solution quality.