Machine Learning Algorithm - Logistic Regression

Machine Learning Algorithm - Logistic Regression

• 1.
  Lesson 12 Machine Learning– Logistic Regression Kush Kulshrestha
• 2.
  Introduction • One ofthe most simple algorithm used for 2 class classification problem. • Logistic Regression can be used for various classification problems such as spam detection. Diabetes prediction, if a given customer will purchase a particular product or will they churn another competitor, whether the user will click on a given advertisement link or not. • Just like Linear Regression uses optimize Least Square function to calculate the coefficients, Logistic Regression uses Log Likelihood function to find the value of the best fit parameters. • Logistic Regression is the basic algorithm used to define a Neural Network.
• 3.
  Sigmoid Function Refresher Thelogistic function or a Sigmoid function is a function which takes any real value between zero and one. It is defined as y = 1 / (1 + e-x) where y is the scaled value and x is the input value. The logistic function transforms the original range of [-,] to [0,1] and also has a linear part on the transform.
• 4.
  Probability, Odds andLog odds Probability of an event = (No of Success / Total no of cases) Odds = (One outcome / all other outcomes) Examples: Conversion: • For rolling a dice and getting 1 – Probability = (1/6) Odds = (1/5) • Rolling a dice and getting an even number – Probability = (3/6) = 0.5 Odds = (3/3) = 1 • Dice roll less than 5 – Probability = (4/6) = (2/3) Odds = (4/2) = 2
• 5.
  What is LogisticRegression Linear regression is continuous response is modelled as a linear combination of the features. In Logistic regression, log-odds of a categorical response being "true" (1) is modelled as a linear combination of the features. This is called the logit function. The equation can be rearranged into the logistic function: p = Conclusion – • Logistic regression outputs the probabilities of a specific class. • Those probabilities can be converted into class predictions. • The logistic function takes on an "s" shape and is bounded by 0 and 1.
• 6.
  Interpreting Logistic RegressionCoefficients Final equation of Logistic regression is: Interpreting beta: ‘1’ unit increase in ‘x' is associated with a B1 unit increase in the log-odds of event for which p is probability of. Log odds are: Hence, positive coefficients increase the log-odds of the response (and thus increase the probability), and negative coefficients decrease the log-odds of the response (and thus decrease the probability).