推移確率に基づく対話型事象検索システム

概要

論文の詳細を見る
This paper deals with the searching mechanism of the event in accordance with the user's aim in a large event set. The pseudo-ordered event set with the probabilistic weights of the relation is introduced,and is formulated as a Markov chain model mathematically. Using the consequent from the probability theory, the event sequence whose events are intensively associated with each other, is easily detected by the computer. In order to deal with the probabilistic model, the interactive Processes are essential. The directed graph, which represents the Markov chain,is always updated through man-machine interactions. Moreover, both the adjacency matrix and the transition probability matrix are modified by the user's response. Finally, user can find out his goal event efficiently. The debugging Process of FORTRAN Programs is also discussed as an application of this event search system.
一般社団法人情報処理学会の論文
1978-03-15