Machine Learning Algo1: Linear Regression Model

Machine Learning online video on Linear Regression by 6benches.in
---------------------------------------------
Machine learning algorithm, Linear regression model(supervised learning) covered in this video lecture.
Training with both feature & target is called supervised learning.
In unsupervised learning, training happens only with features(These algorithms will be covered in upcoming online video lectures)
a) Linear regression concept in theory using a 2 dimensional graph
b) plotting 2 columns using either seaborn or matplotlib
i) seaborn.lmplot(col1, col2, data=dataset)
ii) import matplotlib.pyplot as plt
plt.scatter(dataset.col1, dataset.col2)
c) i)Training of the model
from sklearn.model_selection import train_test_split
x-trainingset, x-testingset, y-trainingset, y-testingset = train_test_split(x-col, y-col, train_size=0.3, random_state=674)
from sklearn.linear_model import LinearRegression
lingmc=LinearRegression()
lingmc.fit(x-trainingset, y-trainingset)
ii) Predicting of the values by model
y_predictValues = lingmc.predict(x-testingset)
iii) Calculating accuracy of the Regression model(r2 score, MSE-Mean square error, RMSE-root mean square error) covered here.
for accuracy of Classification models we use accuracy score, confusion matrix, classification report etc ... covered later
from sklearn.metrics import r2_score
r2_score(y-testingset, y_predictValues)
r2 score gives how good the model is.
iii) Validating model: The other two MSE and RMSE gives how bad the model is.
from sklearn.metrics import mean_square_error
mean_square_error(y-testingset, y_predictValues)

Next lectures will be on
Algo2: Logistic Regression (supervised Learning)
Algo3: KNN (Supervised Learning)
Algo4: K Means Clustering (Non supervised learning)