A Document Classification System Based on Multimembership Bayesian theorem
碩士 === 中國文化大學 === 資訊管理研究所 === 97 === As a result of the development of Internet、make the increasing speed of digital documents faster. So the important of document automatic classification increases too. How to classify the documents quickly and correctly in shorter time is a very important question...
| Main Authors: | , |
|---|---|
| Other Authors: | |
| Format: | Others |
| Language: | zh-TW |
| Published: |
2009
|
| Online Access: | http://ndltd.ncl.edu.tw/handle/88080109220513037882 |
| Summary: | 碩士 === 中國文化大學 === 資訊管理研究所 === 97 === As a result of the development of Internet、make the increasing speed of digital documents faster. So the important of document automatic classification increases too. How to classify the documents quickly and correctly in shorter time is a very important question at the domain of document automatic classification.
In this paper we establish the document automatic classification by Multi-membership Bayesian. It could classify and manage documents usefully. In the training phrase、we establish the database of information word and to compare with the training documents. Finally、we compute the and to the Multi-membership Bayesian formula then we will get the probability of document belong the class. The merit of Multi-membership Bayesian is the value of probability will be modifying by the mount of document increase. We only need to modify some value of opportunity and we will get a new classification module when the new samples get in on. For this reason、the classification module has high mobility. The more documents increase、the more effective module we have.
|
|---|