Abstract
In recent years, the ambiguity of the request definition is concern. As a result, many examples with which a project falls into failure exist. It adversely affected quality, cost, delivery date each targeted for the evaluation of the project, and various actions have been performed. However, it has not resulted in solution of this problem by the requirements analysis method which is in use. Then, data mining is studied as different approach from the requirements analysis which is in use now. It aims at solving the problem that a leak or the hidden demand of user requirements cannot be extracted, by using the data-mining method to the user requirements data of a project in the past. Therefore, to perform the selection of the analysis method was applied association rule mining. As a result, the relationship of various demands was extracted and some unexpected relationship was extracted. A requirement analyst can foresee a user's potential demand from an early stage by using this result from the early stage of the requirements analysis of a similar project. As a result, it can respond also to the addition and change of a requirement which are also the cause of failure of a project.
