HW-0 CS365A

The link of the two different interaction with Eliza are provided below

Reasons for the behavior in A2 -

Instances from the talk -

• Doctor : Are you depressed often?
  Me : no
  Doctor : Why not?

  Here we can see that the response was based on my immediate answer and not on the context of conversation.

  From the code it can be see at :
  (make-local-variable 'neglst)
  (setq neglst
  '((why not \?)
  (($ bother) i ask that \?)
  (why not \?)
  (why not \?)
  (how come \?)
  (($ bother) i ask that \?)))
  

• From the talk one can that it gives "depressed" a highest priority and whenever it runs out of key word it goes back on asking various questions on depression, like :

  • Doctor : I understand. So what makes you think this is so?
    Me : you are talking no sense to me.
    Doctor : Why do you say you are depressed?

  • Doctor : Is it because of problems at school that you say you m done?
    Me : why school ?
    Doctor :What do you think?
    Me : i think i should go now.
    Doctor : Earlier you said you felt depressed?

Suggestions :

• History and the context of the conversation must be given proper weightage.
• Can use some machine learning techniques and train it with some real conversations.