Preprocessing Data for Machine-Learning Algorithms to Provide User Guidance in Special Purpose Machines

摘要

In our previous work, we proposed a system which makes complex production machines more user-friendly by giving the recommendations to the operator. So, we assist the user working with a complex production machine. The recommendations are presented like: "In the last 10 occurrences of this event the operators performed the following keystrokes". While working on the project, we had problems with retrieving the correct recommendations from our knowledge base. Meanwhile, we gathered more data from our project partners. Now, we dive deeper into this data in order to improve our solutions. This work describes methods to preprocess the data. This preprocessing should help us building up the knowledge base. To achieve this automatically, we do not want to know much about the machine and the production process itself.