Since the beginning of chatbots, the most common techniques used for simulating conversation
has been different pattern matching-techniques. This is a study of possible alternative methods
and an attempt of implementation. As the research of natural language processing is rapidly
progressing, the aim of this report is to clarify current techniques and what could possibly be
achieved by adapting recent findings on the area.
A combination of part-of-speech tagging, methods for information retrieval and known methods
for generating language are covered and described of how it could be of use. An implementation
is reviewed and compared to traditional programs. The final conclusion states that the
implementation is not as satisfying as chatbots using elaborated pattern matching-techniques.
However, it is concluded that the new methods covered are more likely to be part of future
chatbots considering the progress of natural language processing.