Abstract
Most project-based organizations possess extensive collections of diverse project documents. Exploring the knowledge codified in such project documents is specifically recommended by the common project management guidelines. In practice, however, project managers are faced with the problem of information overload when trying to analyze the extensive document collections. This paper addresses this problem by combining two approaches already established in other disciplines. The first involves the development of a Project Knowledge Dictionary (PKD) for the automated analysis of knowledge contents codified in project documents. The second involves the integration of a sentiment analysis where concrete opinion expressions (positive/negative) are identified in connection with the codified project knowledge. Building on this, three mutually complementary analyses are demonstrated, which provide the following contributions: (1) determining the volume and distribution of five project knowledge types in project documents; (2) determining the general sentiment (positive/negative) in conjunction with the textual description of the project knowledge; (3) classifying project documents by their sentiment. By this means, the proposed solution provides valuable insight into the emotional situation in projects and contributes to the emerging research issue of project sentiment analysis. Furthermore, the solution makes a contribution to overcoming the information overload by assessing and organizing the knowledge content of large document collections.
Recommended Citation
Matthies, Benjamin, "FEATURE-BASED SENTIMENT ANALYSIS OF CODIFIED" (2016). PACIS 2016 Proceedings. 144.
https://aisel.aisnet.org/pacis2016/144