majdialali/AI-chatbot-for-Norwegian

In this big-size master-level project, we are going to build an AI chatbots for Norwegian language. The main goal of this chatbot is to answer on general questions of who are learning Norwegian as a second language. It is meant to help them make some natural conversations. But as the data grows and varies, we can use this chatbot for other purposes like for industrial uses. To build this chatbot, we are going to use a neural network together with some Information Retrieval techniques and NLP tools. More specifically we will create a neural network which can classify the user input and return a response back to the user. As a part of the preparing of the data, we will use NLP-tools like lemmatizing. At the end of the preprocessing stage, we are going to use two approaches to numerically represent our data. The first one is Bag of Words, an IR technique, while the second one is called tf-idf, term frequency-inverse document frequency, an NLP-algorithm. We will also do some adjustment on the user input, when necessary, to optimize the model where we will make a spell checker from scratch. This spell checker is capable to detect the misspelled Norwegian words and suggest words to substitute them. This can give better perdition of the classes. Consequently, it returns qualitive responses to the user.