What is Regression , Maths or Machine learning ... awwww

In general, regression is defined as a technique to find the relation between two events/variables/activities.  

For Example:

  • The relation between hard work done by students throughout the year to salary package they get during the campus interviews.
  • The relation between the size of the house to the price of the house.


                  



If you remember your schooldays with 2-D coordinate geometry  (Thanks to S L Loney). He says the equation of a line is  "y=mx +c"

  • y: Coordinate on Y-axis
  • x: Coordinate of X-axis
  • c: Intercept (where the line will intersect at Y-axis)
  • M: Slope of a line. You can understand how much line is tilted towards X-axis. It is a number that measures lines "steepness" also called the gradient of a line.
    • Slope of line is calculated as   (y2-y1/ x2-x1) , here (x1,y1) and (x2,y2) are two points on line in 2-dimension.


  


 So if you know x,m, and c you can find y .i.e y is dependent on the value of x, if x will change y will also change. "If you do more study, salary package will be more"

In statistical terms, Regression is a technique used to model and analyze the relationships between variables and oftentimes how they contribute and are related to producing a particular outcome together. Regression techniques are used when the output is real-valued based on continuous variables. For example, any time-series data. This technique involves fitting a line. Regression can also be thought of as looking for a trend line in our data. When used in machine learning, that line allows us to predict a numeric value based on features. We could use regression to predict things like:

·         How much money something will cost
·         What's the expected temperature tomorrow
·         How many days until you need to replace your car battery


It is a form of predictive modeling technique that investigates the relationship between a dependent (target) and the independent variable (s) (predictor). In the Regression problem, we try to predict results within a continuous output, i.e. we try to map input variables to some continuous function. (y=a+b*x)

Regression majorly depends on
  • Number of independent variables
  • Type of dependent variables
  • The shape of the regression line


Types of Regression:

  • Linear Regression
  • Logistic Regression (it is considered as classification)
  • Polynomial Regression
  • Stepwise Regression
  • Ridge Regression
  • Lasso Regression
  • Elastic Net Regression


Benefits of Regression:

  • It indicates the significant relationships between the dependent variable and the independent variable.
  • It indicates the strength of the impact of multiple independent variables on a dependent variable.


General Regression Problems:

  • Given data about the size of houses on the real estate market, try to predict their price. Price as a function of size is a continuous output, so this is a regression problem
  • Given a picture of a person, we have to predict their age on the basis of the given picture
  • This technique is used for forecasting, time series modeling and finding the cause and  effect relationship


No comments:

Post a Comment