Discovering Fails in Software Projects Planning Based on Linguistic Summaries

Linguistic data summarization techniques help to discover complex relationships between variables and to present the information in natural language. There are some investigations associated to algorithms to build linguistic summaries. But the literature does no report investigations concerned with combination linguistic data summarization techniques and outliers’ mining applied to planning of software project. In particular, outliers’ mining is a datamining technique, useful in errors and fraud detection. In this work authors present new algorithms to build linguistic data summaries from outliers in software project planning context. Besides, authors compare different outliers’ detection algorithms in software project planning context. The main motivation of this work is to detect planning errors in projects, to avoid high cost and time delays. Authors consider that the combination of outliers’ mining and linguistic data summarization support project managers to decision-making process in the software project planning. Finally, authors present the interpretation of obtained summaries and comment about its impact.