Predicting with sparse data

It is well-known that effective prediction of project cost related factors is an important aspect of software engineering. Unfortunately, despite exte nsive research over more than 30 years, this remains a significant problem for many practitioners. A major obstacle is the absence of reliable and sys tematic historic data, yet this is a sine qua non for almost all proposed m ethods: statistical, machine learning or calibration of existing models. In this paper, we describe our sparse data method (SDM) based upon a pairwise comparison technique and Saaty's Analytic Hierarchy Process (AHP). Our min imum data requirement is a single known point. The technique is supported b y a software tool known as DataSalvage. We show, for data from two companie s, how our approach-based upon expert judgement-adds value to expert judgem ent by producing significantly more accurate and less biased results. A sen sitivity analysis shows that our approach is robust to pairwise comparison errors. We then describe the results of a small usability trial with a prac ticing project manager. From this empirical work, we conclude that the tech nique is promising and may help overcome some of the present barriers to ef fective project prediction.