Applying the Random Forest Algorithm to Predict Engineering Change Effort

Shorter development cycles and increasing product complexity are a major challenge within the product development. Furthermore, more engineering changes are taking up a high amount of development resources. Therefore, the identification of effort for engineering changes is important to identify necessary resources. This paper introduces a methodology to build a model to predict engineering change requests' effort. Therefore, first a description model for engineering change is introduced. Afterwards, existing engineering change data has to be transferred into a data set by use of the description model. The build data set is split into training and test data to identify the prediction quality for unseen engineering change requests. Next, the prediction model is built in python by the aid of Sklearn, NumPy and Pandas. The results of the prediction model are presented and discussed. Furthermore, the conclusion suggests further possibilities to improve the prediction.