Text mining for incoming tasks based on the urgency/importance factors and task classification using machine learning tools

In workplaces, there is a massive amount of unstructured data from different sources. In this paper, we present a case study that explains how can through communications between employees, we can help to prioritize tasks requests to increase the efficiency of their works for both technical and non-technical workers. This involves managing daily incoming tasks based on their level of urgency and importance.To allow all workers to utilize the urgency-importance matrix as a time-management tool, we need to automate this tool. The textual content of incoming tasks are analyzed, and metrics related to urgency and importance are extracted. A third factor (i.e., the response variable) is defined based on the two input variables (urgency and importance). Then, machine learning applied to the data to predict the class of incoming tasks based on data outcome desired. We used ordinal regression, neural networks, and decision tree algorithms to predict the four levels of task priority. We measure the performance of all using recalls, precisions, and F-scores. All classifiers perform higher than 89% in terms of all measures.