Make a Chatbot in Python (Make it Do Anything) | Intro to Data Science

[@SATSifaction]

If you like this content be sure to comment like and subscribe. Be sure to check out how to build a chatbot with Rasa for generation of a conversational AI bot 👉🏻 kzbin.info/www/bejne/ZqCxXq2NhZ54oLs Be sure to check out my machine learning for beginners playlist here 👉🏻 kzbin.info/aero/PLM30lSIwxWOihE6bxSuSSzJcFoinq5FPu

[@codewarrior4458]

Thank You

[@codewarrior4458]

I have never come across an explanation on Chatbot like this

[@creativecore3575]

Thank you so much for breaking down what Data Science means using your own created example

[@creativecore3575]

Hi! I was really curious about how you managed to leverage the Covid 19 API. How would you apply it in a different country ? This project is so cool !

[@ansgodupdate9773]

i had an error trying to execute the train.py file. this is the error ModuleNotFoundError: No module named 'pyty' plz help me to solve it

[@artistpw]

Unfortunately, chatterbot only works with python from versions 3.4 to 3.8 from a stackoverflow posting.

[@SATSifaction]

I think chatterbot is now the older way of doing things my recommendation is to build a small bot using an LLM and fine tune it with rag

[@theharawijayaratne3088]

What python version were you using ?

[@akarsh704]

This might be a stupid question, But does train.py also use the generated sqllite3 file for training?.. if not give some tips on how to convert the database into a .yml file so i can use that to train the bot next time? Its a bit confusing for a noob like. Great video! 😇

[@SATSifaction]

Great question. No it uses the yaml file the database is there for you to reference actual answers and questions people are asking. You could write a script that pulls questions from the db that are not in your yaml corpus and suggest how to answer them based on your product/service/requirements.

[@akarsh704]

@@SATSifaction thank you! 😇

[@ssprajapat3135]

hi there, great video thanks, please keep making more tutorial like this, thanks

[@ashiqjohns6375]

Hi Sir...would like to clarify a doubt. Hope you would help...

[@SATSifaction]

What’s the doubt

[@suryadevsudheer5947]

can someone tell me how to properly run this code?

[@jahmaijones3353]

Where did you get the English words corpus from ? And how to import into Python

[@joecool8460]

Do you have a version with python over Django framework?

[@SATSifaction]

I dont but its a great idea for a future video

[@PhotoSlash]

the problem is that this is nothing new, making an intelligent bot using just a dataset of pre-existing conversations doesn't make it smart, it just repeats other people's answers to your questions, therefore you might and will get wrong or outdated answers. I still can't find a valid course on how to develop a smart assistant which generates valid and updated answers, you can use datasets for specific groups of interests like food, animals, movies, etc where you know that their datas are valid also in the future and implement them in a response of the bot. Using a dataset for generic conversations is kinda pointless nowadays, sure you might find useful and smart answers but not 100% reliable.

[@SATSifaction]

I dont think you quite grasped the proof of concept here. Towards the end I talk about new questions that are outside of the realm of your generic hello or hi....those new questions need to be trained with responses...if you let the AI loose and train its own answers without human input to guide it you will end up with what MS had telling people to harm themselves. So in an effort to be ethical, you should always review what your bot is responding. Furthermore concepts such as reinforcement learning is used often to train models such as this case. You extract all the questions without suitable answers, run ML on them and determine the best answer....

[@lwazidlamini3440]

Any developer that can deliver such a simple product to connect to an API? Custom development

[@AKSHAY-ib9to]

can i get the source code

[@digvijaysingh9613]

raise IOError(Errors.E053.format(path=path.parent, name="meta.json")) OSError: [E053] Could not read meta.json from en

[@digvijaysingh9613]

I am getting this error while executing spacy.load('en')