Abstract
In order to improve efficiency for suggestion of data analysis approaches, this research develops a problem-oriented model for suggestion of data analysis approaches. Firstly, this research analyzes the historical problems the applicable data analysis approaches and other information of the original solutions which are related to data analysis. Secondly, a model is developed for analyzing the solutions related to data analysis and transforms them into structured forms. After that, the proposed model can be applied for searching related data analysis solutions based on the problem characteristics. In order to distinguish the related and the unrelated solutions, this model evaluates the likelihood between those solutions by the linear combinations of the solution characteristics. Further, this model utilizes the simulated annealing algorithm to acquire the combinations which can distinguish the related and unrelated solutions. Finally, this model can suggest the data analysis approaches according to the approaches used in the related solutions used. As a whole, the proposed model can assist managers realizing data analysis approaches in an efficient and effective way.