Model design and development for user support system using artificial intelligence techniques in enterprise resource planning software

From the point of view of organizations, it is very important to manage business processes correctly, consistently and transparently. The management of processes is also related with the proper management of data. Thus, this is possible with enterprise resource planning (ERP) software designed around a central database, where all departments can work integrated. Users are one of the most important factors in the success of ERP software for the organization. Due to many reasons such as lack of training and experience, technical problems, users need support during the use of ERP software. This support, whether it is provided from internal or external sources, causes a cost. This study aims to find an answer the question as to whether a structure can be developed in which users can resolve their support requests themselves. For this purpose, an intelligent user support system model with an artificial intelligence-based process engine is proposed for ERP users. With this model, unstructured user support requests created with natural language are structured with natural language processing methods and classified by multi-class machine learning algorithms. It is ensured that the process belonging to the relevant class is operated and the users can produce solutions to their own problems. Moreover, 1000 text samples written in natural language in 10 different categories based on real data were used in the learning process. Different term weighting methods were applied after the natural language processing methods of the data. Then, it was analyzed with machine learning algorithms (K-Nearest Neighbor, Naive Bayes, C4.5, Support Vector Machine, Random Forest, Sequential Minimal Optimization, LibSVM) and deep learning algorithm (Long Short - Term Memory (LTSM). The best classification was obtained by Sequential Minimal Optimization (SMO) with TF-IDF weighting. In the application development phase, an intelligent support robot was developed with the SMO classification algorithm. Thanks to the process engine of the support robot, user support requests are directed to predefined flows. In addition, this section includes the use of virtual reality and augmented reality technology to provide human-computer interaction in the user support system. This thesis contributes to the field in terms of both the development of a support robot that focuses on the problems of the users and the technologies used in this process. Keywords: Enterprise Resource Planning, User Support, Artificial Intelligence, Virtual Reality, Augmented Reality, i-ERP