Notes of the meeting with T.A.(Customer of our project) (20.02.2017) - bounswe/bounswe2017group7 GitHub Wiki

Meeting Note: 20 Feb 2017

Attendees: Irmak,Ebru,Cagdas,Salih and Arda Celebi(T.A. and customer)
Place: BM AVS room.

Guidelines

  • Our objective for this meeting is to clarify project specifications.
  • We will discuss our requirements that we decide on last meeting.
  • Questions from the last meeting will be asked to T.A.

Notes

  • To use and implemet API's there are some recommended videos and tutorials on the net. T.A. will send them via email.
  • We are going to be mostly dealing with Intelligent Response part of the ChatBot
  • Recommended programming languages are Python and NodeJs. Functional programming languages may be useful for filtering but not recommended.
  • Since we are going to use Telegram API, there is no need for two different accounts(user and guest). It is enough that if user has a Telegram account.
  • NLP tools may ease our work but they have low response time and not recommended.
  • We can use some tools for typo mistakes that user might do.
  • Our job is to match the user input with the templates in our databese.
  • To increase functionality we can do this matching process within two levels. First level finds the spesific part of the input and matches to a template. Second level tries to divide the first template into two or more different templates and tries to match them with the information in the database.
  • MongoDB has been recommended.
  • We don't need to match every possible user input with the templates.
  • The idea searching any book with their famous quotes is not useful but we can use famous quotes as an extra information about books.
  • Creating profile page for each user is not necessary.
  • We should add features which are going to be used. Spending time on implementing non-useful features doesn't make any sense. We should review our requirements according to this measure.
  • We can do a conversation graph for Intelligent Response part.
  • We can use Bot Analytics in order to find which node has been visited at most in our conversation graph.

TODOs

  • Share all information above with the group members.