Applying Modern Text Processing Technologies to Implement a Self-learning Marketplace by Using Cloud Services as an Example

Marketplaces have become common ways to provide systems for commercial interactions between users and vendors. These systems must be maintained and supported by people adapting the system to the user’s behavior. Users behavior can be tracked by analyzing the text inputs users do while they search for a product. The goal of this research was finding a way to automate this adaption process with considering the inputs of the users on a marketplace prototype for cloud services. To fulfill this task, modern methods of text processing such as stemming, and lemmatization were suitable. The problem has been solved by applying a self-learning algorithm combined with a self-updating database. The system now provides direct feedback recognizing wrong input. As a result, future research could be the further development of the self-learning system to a self-adapting system.