Tamagotchi

Inspiration

I wanted to make a chatbot in Tensorflow but just that wouldn't make the project special. As a result, I made a little Tamagotchi game in Java where you can talk to your pet.

What it does

Right now, numbers just go up and down when you interact with your pet. I followed a tutorial on how to make a chatbot with to learn and the code for that is in another repository.

How I built it

Tamagotchi

• two Java classes: pet and Tamagotchi
• play from Tamagotchi

Chatbot (I just followed the tutorial)

• Made a JSON file with all the tags, questions, and responses
• Break each sentence into words using stemmer, then break down the word again
• All the words are added to a list called 'words'
• Convert all sentences to 1 if the word is in 'words,' else converts to 0.
• Create the neural net layers and train the data

Challenges I ran into

• I started off trying to do sentiment analysis using Doc2Vec but that didn't turn out well
• I attempted to train my neural net on a dataset with a lot of movie subtitles but that didn't work out either
• Trying to stay awake while not binge eating

Accomplishments that I'm proud of

• The chatbot learned the data I wrote!! Although, you can't chat with it yet.

What I learned

• What API is (was looking up how to implement Python code into Java)
• Neural network structure
• How to train a neural network
• basic tflearn
• What intent is in Chatbot Design
• A little bit of DialogFlow

What's next for Tamagotchi

• Rewrite the chat bot part since I pretty much copied most of the codes
• Add time to the game
• Make it possible to chat with the chat bot

Built With

• java