DAY17.md

DAY 17

Git Hub Branching

ubuntu 18.04 on aws and Jupyter Notebook

• git branch number of branch

• git checkout -b name creat new branch

• git log show all commits

• git reset --hard commit-id to reach the first repo or go to specific commit

• git push origin branch

• git checkout master <--- branch name change the branch

Supervised Machine Learning

data---> computer--->1st---> predict,question

• We need a DataSet for training

• Then we give this data to system for training for learning purpose

• Then the training model is ready for QnA.

• Then we give a new question related to the training model.

Example

• When we see a dog for the first time, someone introduce us that this object is a dog. Similarly when we see another breed of a dog or an animal we can identify that which animal it is.

• The actual object is called as label

• The attributes | Features | Characterstics of an object are features for computer.

• So we have to give both label and its features to the system for training purpose.

• Model Data

• Apple Texture : Smooth Weight : 100-120 Grams

• Orange Texture : Bumpy Weight : 120-140 Grams

• Here Smoooth is for apple and Bumpy is for Orange

features=[[100,"Smooth"],[120,"Smooth"],[130,"Bumpy"],[150,"Bumpy"]]

label = ['apple','apple','orange','orange']

• Now we need an algorithm for training purpose.

Supervised Machine Learning

• Classifier This means when we need to distinguish between features and attributes or basically classifying data

Data Classifiers techniques for Supervised Learning

1. KNN
2. SVM
3. Decision
4. Naive Bagos
5. Random Foresh
6. Lion
7. Me
8. Regression will discuss later for this.

• In python there is a liberary called scikit which contains all the above classifiers. It is a framework containing all the classifiers written in python.

• Since by default we dont have scikit-learn installed so we have to install it with pip3 pip3 install scikit-learn

Decision Tree Classifier for Supervised Learning.