Read Instructions Please! <-- I tried to use a ML approach to make predictions for Pico's emotions. In this "model", you "train" the project to become smarter and to accurately predict Pico's emotions based on a color (a catalyst to trigger emotions in context of this project). Click Generate! "Generate" will generate a random "catalyst" for Pico's emotions. The "catalyst" is the colored box that shows "I cause _____" on it. Most likely, our "model" will incorrectly predict the emotions (example: catalyst shows green but Pico was predicted to be angry). The color key at the bottom shows the correct emotions Pico should show based on each color box/catalyst. Pico's emotions are determined "randomly", but as you click on "generate" more, it will realize that what emotions are "False" and which ones are "True". Check code inside for more details. As you generate more, or "train" the model more, the accuracy will increase. Of course, this "model" is very inaccurate as its made using Scratch Lists and there's only 4 lists/emotions to decide between. Notes: - The process is classification of supervised learning since it uses labeled i/o data (catalyst/emotions) and it "classifies" the "data" into various (4) "classes" but idk. - If the data was generated randomly and automatically using a pre-written function, also automatically using various models to "train" the model, the process would be similar to AutoML.
Please heart/star the project if u thought it was interesting! | 请点个赞吧(关注会更好哦)! | 这个项目不能完全说是“AI”,我只不过用了ML训练模型用的过程而已。 : ) This project isn't totally "AI"; I only simulated the training process ML uses. In reality, I believe the data needs to be much more and we must import and apply various models to determine which holds a higher accuracy. Also, we would have X and Y variables for each testing and training data. |'`~ note how i didn't say AI, i said "AI" ... im still exploring this around too :P