We have been assigned a project, in which we have to create a chatbot which will ask question, take the replies, analyse them and give an approximate assessment of the current emotional state of the person. There are two aspects of the project,

  1. Training the bot to choose the next question based on the previous response
  2. And analysing the responses individually, to detect . the sentiment.

What technology would we have to use and what would be the steps to accomplishing the tasks?


  • $\begingroup$ A Socialsent sentiment lexicon can formalize opinions in a model. This is a starting point for training a chatbot. $\endgroup$ – Manuel Rodriguez Mar 3 '19 at 12:44

There are some different approaches to learning this sequential acting. First, you can use RL (reinforcement learning) and define some rewards over the action of the user over that question, to explore and exploit the environment.

Second, you can use RNN to predict the preferred next question from a history of data. In contrast with RL, you need many instances to train the network to predict better next answer more accurately.

| improve this answer | |

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.