Please enable JavaScript.
Coggle requires JavaScript to display documents.
2.4) How to train and Assess ML Model, 3.2) The line of best fit, 3.3)…
2.4) How to train and Assess ML Model
model training
Process of educating the model about the relationship between the predictor and the response variable
The model preform mathematical calculation and preform parameter adjustment using optimization algorithms
Train-test split
The data we poccess will be split randomly[preventing any bias ] into 80% - 20% ration
Assessing the model performance
Method will depend upon the problem being solved
Regression metrics
Mean Absolute Error {MAE}
Mean squared error {MSE}
R-squard
Classification metrics
Prediction
Recall
Accuracy
F1 Score
3.2) The line of best fit
Define it
It minimize the overall distance between the line & Data point [line is as close as possible to al the point ]
Residual /Error
The distance between each data pint & the Line
Tool to visualize the relationship between to variables to identify a trend, hence predict future data points
How to use the linear regression ?
Make Predict for dependent variable based on the existing data
determine the relationship between the variable
Determine the strength of the relationships
Correlation coefficient
The measure of the strength & direction of the linear relationship between the 2 variables
Range between -1 to 1
Features
It decrease the sum of square residual [FOR ALL DATA POINT ]
Residual is squared to ensure all value are positive as there is points above & below the line
The line of best fit minimize the sum of squared residual FOR all data points
3.3) Line of best fit Equation
Define the equation
y=mx+c
c
y-intercept
point at which the line intersect with the y-axis
value of y when x is zero
WHAT IS ITS IMPORTANCE ?????
x
independent variable
m
slope of the line
indicate the rate of change in y per unit of change in x
steeper slope= more significant impact of x on y
sign of m
indicate the direction of the relationship
Accuracy of the predicted values is important to determine the correlation coefficient
The closer the predicted value to the actual value ,The higher the correlation coefficient = Actual value lie close to the line of best fit
Use the equation to make prediction
Equation Tell us the what value we expect to for y based on the value of x
When we use it?
What is linear regression ?
Technique to find the best position for the line of best fit where the sum squared residual is at its lowest possible value
Statistical method that allow us to analyze the linear relationship between dependent and 1/more independent variables
it involve fitting the straight line to a set of data points to describe the relationship