Hey guys ;)
So as the title said I’ve finished my Music Classification project. You could check it here Music Classification Project.
Sorry that I didn’t deploy this app. I hope if someone will be interested in it, then this person will just copy my repo.
What can I tell you?
Actually I don’t know, this project was pretty easy, because only librosa was new for me. All the other libraries I’ve used before. Maybe I just should say the main points of how to transform data into models. I guess that will be good.
First step
Maybe it will be unexpected, but you should choose a topic in which you’re interested. I think it’s important for every project you’re working on. If you don’t have fun with it, then why are you doing it?
For example all my projects built on my interests and hobbies.
Second step
Now, you need data. It’s good if you will find an amazing dataset for your topic, or you just could scrape all data manually. The more proper data you have, the more accurate your model will be. I mean if we only have 10 students for example and half of them are good students and the other half are bad, then how to classify an average student?
Third step
Here you need to work with data I guess. Clean it, fill None values, normalize data and split data into training and test datasets. That’s easy, maybe normalization is a bit difficult, because you should choose for example between StandardScaler and MinMaxScaler. Also for some models you need to encode target.
Final step
Choose the right model and make it accurate. Yep there are Regression models and Classification models, and there are Linear models, ensemble models and so on. How to choose right? You need to read a bit of theory.
What’s next?
Again the title said everything already. I’m going to come back to my Aster - personal assistant project. Actually I had 3 options - Chess Project, Aster and brand new AI for Cuphead.
But I think I need to work with Aster, at least upgrade it, and it would be great to make Aster “alive”. So, for the next few weeks I’ll be writing about Aster. That’s all for today, see you guys later
