I Recently joined ML Zoomcamp to build my ML skills.
These are some things I have learnt, I hope you could learn something too.
Week 1:
- What's Machine Learning ?
Machine learning is the ability of machines to pick patterns from data and its features, and use the patterns to make predictions to new data.
g(x) ≂ y
- ML VS Rule-Based Systems(RBS)
RBS systems involve you identifying every possible scenario and creating a rule for such, this is a very tedious process and is very limited and time-consuming, these rules can instead be used as features to a ML model. The model is trained on a collected data with these features.
- Supervised ML
In this type of ML, a model is trained on features (x) and their corresponding target variable (y) or labels.
- CRISP-DM (Cross Industry Standard Process for Data Mining)
CRISP is a methodology for organizing ML projects, it's been around for a while now, old but still used. It consists of 6 major steps as shown below..
- Model Comparison Problem
When comparing models in your evaluation, be careful, as models can get lucky, so it's advisable to also perform validation and test on each of the models you're comparing to make the best choice.
Train -> Validate -> Test
Also had a numpy, pandas and linear algebra refresher. Then was given an assignment to reinforce all this.
You can also follow via this repo Ugo's ML Zoomcamp
Top comments (0)