2010 年 27 巻 4 号 p. 316-326
It is important for software engineers to improve their skills by self-learning existing software development techniques. However, it is not easy for engineers to find out which techniques should be learned because there are so many existing techniques and also new techniques are emerging continuously. To support software engineers’ self-learning, we propose a development technique recommendation method using collaborative filtering (CF). In our method, a user (engineer) gives ratings to development techniques based on his/her interest or benefit, then development techniques are recommended based on similarity measure of CF. The similarity is computed based on co-occurrence of terms of development techniques on web pages. We conducted an experiment to evaluate the recommendation accuracy of the proposed method. As a result, the average NDPM was 0.41 for recommending development techniques that are unfamiliar to software engineers.