As a project, I thought I try to see if my assistant could learn, or at least understand to some extent emotions.
I currently have 2 word arrays, Positive and Negative.
When the user inputs a response to "How do you feel?" the input will be scanned with the two word lists.
For each Positive word found, a variable (emotion) is + 1, and for every Negative word found, its -1.
I've used this method to get a scale of -10 to +10. of which I can form a response depending on the number.
I've been trying to work on responses for each number, but giving an appropriate response for the scale can be a challenge.
But using this, means that my assistant could learn what mood I'm in depending on the input I give.
Further improvements could be multiple random responses depending on the scale.
This is to say each scale will have multiple possible responses but it'll give a random response.
What responses do you think could be appropriate?
No comments:
Post a Comment