Content area
Full Text
Abstract
Recently, the interest on the personal information is increasing, and relevant policies and technical development are performed, leading to emphasizing the importance of personal information protection. In this study, the emphasized personal information protection policy was applied to study the inference method for user's personal identification based on the limited personal information. The user's feature vectors were extracted, and the Bayesian Inference Method was applied on this to verify the personal identification correlation. The data that was not studied through the Bayesian Method was applied with the smoothing processing method to maintain the stability in the inference probability. Moreover, through this personal identification correlation verification, it is expected to also perform the role of disabling malicious illegal use or handling.
Keywords:, Cognitive Rehabilitation, Classification, personal information protection
(ProQuest: ... denotes formulae omitted.)
1. Introduction
Along with the development of the ΓΓ technology, great amount of information are expected to be search faster and more accurately. The technologies in information search include keyword base, cross language search and automatic classification, but these methods all provide text sets such as documents or paragraphs as the processing result On the other hand, the Q&A type search technology first identifies the intention in the user's question, and finds the answer from the documents subjected for search to be provided[l][2]. To extract the answer from the document the method of recognizing and extracting the entity name such as a person's name, organization name, name of the place and unit is required [3]. In this study, the features of the Korean name and the American name among the English-speaking world are compared and classified to be studied, and titen a specific name is inputted to verify the gender on the name. For recognizing a person's name and gender distinction, the features of the syllable that are shown in the last and first name, and the statistical information are used [4]. For the names of Koreans and foreigners, the Naive Bayesian classification feature was applied.
2. Related Works
2.1. Bayesian Model
The Bayesian Model is based on stochastic approach, therefore, tie parameter of the model is resided as random cæ fficient In the Bayesian Model, it is assumed by probability distribution. This probability distribution can indicate the uncertainty that is...