Machine Learning Tutorial

Hello everyone, Today I'll tell you about how we can code to teach machine to learn from experience.
We will use python in Ubuntu 16.04.
So, lets get started. There is mainly 2 open source machine learning library,
1. Scikit Learn
2. Tensorflow

We'll use scikit-learn here
So, to install scikit-learn using pip, you have to execute this command-

pip install -U scikit-learn

to install scikit learn using conda

execute this--   

conda install scikit-learn

after installing scikit learn test it if it's installed correctly or not.

write 'import sklearn' in test.py file and save it.

after that compile it, if there is no error then congratulations. you have successfully installed scikit learn.

Now, we will look at our training data and testing data
To train your ML model you need the training data and to test you need testing data.

there are lot of sample datasets available to test.
here is the link- https://archive.ics.uci.edu/ml/index.php
Iris dataset is one of the most popular dataset, you can start with that.

In Iris datasets there are description of 3 types of iris flower. we will predict the flower type just by sepals and petals length of flower.

First of all we've to add the libraries we need
Do as follows -

import numpy as np

from sklearn import datasets

from sklearn import tree

# load the iris datasets

dataset = datasets.load_iris()

#Keeping 3 record a side for testing purpose

#index 0 indicates the first record of Setosa, Index 50 for Versicolor and 100 for virginica

testing = [0,50,100]

#deleting records that kept for testing

train_target = np.delete(dataset.target,testing)

train_data = np.delete(dataset.data,testing, axis=0)

#storing 3 data for testing

test_target = dataset.target[testing]

test_data = dataset.data[testing]

# fit a CART model to the Training data

model = tree.DecisionTreeClassifier()

model.fit(train_data, train_target)

#checking the data we stored for testing

print(test_target)

#testing if the model can recognize this set of data it has never seen

print(model.predict(test_data))