- Aug 6, 2012
- 27,986
- 24,787
- 2,405
I'm more interested in Machine Learning and A.I. In fact, I felt this project (still not complete) was taking me way from my other interests, but, it seems having a portfolio of projects is vital. I am interested in Natural Language Processing (NLP), so it's "sort of A.I" type of endeavour which allowed me to better understand some key NLP modules. I have many ideas for a chatbot that I haven't seen implemented by anyone else, so there are possibilities.
Anyways, for any python nerds, I used NTLK (natural language toolkit), sklearn and a few other minor modules. My challenges now are linking the chat to the GUI I built, since I'm not a good enough programmer to have coded them both together, they are separately coded features. So currently the chat is in my IDE console. I'd also like to save the data into a local SQL DB to try some different true, bot A.I (rather than the rules-based approach I took).
In essence, I used ntlk module to extract information from a covid-19 article and use the context of my question to feed the response from the bot. I had set up some shutdown responses (so, if I said "thank you", the bot would say "goodbye" and break from the code), but funny, I type "ok" before thank you and received a second unexpected response.
So here are my first two sentences and the bot responses based on the article:
Health Expert: I am a Health Expert, I will answer your questions about Covid-19.
Are we going to get out of lockdown soon?
Health Expert: If R remains one, we get a plateau.
ok thank you
Health Expert: We’re now all very aware that if you’re tracking cases, hospitalizations or deaths epidemic, they roughly follow a bell-shaped curve.
So those are my first two sentences, I have more now as I mess around with it. I'm quite proud as I'm not a software developer and I am self taught, the possibilities are truly endless. Anyone can do it if they have an interest. The field is so deep.
Anyways, for any python nerds, I used NTLK (natural language toolkit), sklearn and a few other minor modules. My challenges now are linking the chat to the GUI I built, since I'm not a good enough programmer to have coded them both together, they are separately coded features. So currently the chat is in my IDE console. I'd also like to save the data into a local SQL DB to try some different true, bot A.I (rather than the rules-based approach I took).
In essence, I used ntlk module to extract information from a covid-19 article and use the context of my question to feed the response from the bot. I had set up some shutdown responses (so, if I said "thank you", the bot would say "goodbye" and break from the code), but funny, I type "ok" before thank you and received a second unexpected response.
So here are my first two sentences and the bot responses based on the article:
Health Expert: I am a Health Expert, I will answer your questions about Covid-19.
Are we going to get out of lockdown soon?
Health Expert: If R remains one, we get a plateau.
ok thank you
Health Expert: We’re now all very aware that if you’re tracking cases, hospitalizations or deaths epidemic, they roughly follow a bell-shaped curve.
So those are my first two sentences, I have more now as I mess around with it. I'm quite proud as I'm not a software developer and I am self taught, the possibilities are truly endless. Anyone can do it if they have an interest. The field is so deep.
Last edited:
