Very important to resolve requests or complaints by directing the demands and complaints submitted by the customers to the correct business units in corporate companies. Automatically interpreting and transmitting such incoming information and texts according to the relevant business unit category eliminates a huge human resource burden, saving time and resources. In this study, it is aimed to make sense of the information in the forms that come to a corporate firm as a customer complaint or request and to convey it to the business units divided into certain categories. A system was tried to be developed to determine in which categories the files may be based on the words or the content of the texts by scanning the file. An automatic classification system was developed by evaluating the incoming information on a pre-trained data set belonging to certain categories. In this classification process, using the TensorFlow, Keras, and Torch libraries, which are among the deep learning libraries, with the help of the BERT model, the machine learning algorithms were tried to he automated with Word2vec. By using data mining pretreatment algorithms on the document or form in the study, the data is transformed into a workable structure. Keywords are weighted with this generated study data. Finally, with the help of machine learning algorithms, the text is classified according to the specified categories. The success of the obtained classification results is shown in detail.
