Scribd Logo
Upload

Welcome to Scribd!

  • 9 views

    Session 15-Logistic Regression

    Uploaded by

    Pratyusha Voruganti

    Copyright:

    © All Rights Reserved
    Download as PPTX, PDF, TXT or read online from Scribd

    Session 15-Logistic Regression

    Uploaded by

    Pratyusha Voruganti
    9 views16 pages

    Copyright

    © © All Rights Reserved

    Available Formats

    PPTX, PDF, TXT or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    Copyright:

    © All Rights Reserved
    Download as PPTX, PDF, TXT or read online from Scribd
    Download as pptx, pdf, or txt
    9 views16 pages

    Session 15-Logistic Regression

    Uploaded by

    Pratyusha Voruganti

    Copyright:

    © All Rights Reserved
    Download as PPTX, PDF, TXT or read online from Scribd
    Download as pptx, pdf, or txt
    You are on page 1of 16

    Logistic Regression

    Dr. Rajiv Kumar


    IIM Kashipur

    Note: Content used in this PPT has copied from various source.
    Introduction

    Logistic regression is a method for analyzing relative probabilities


    between discrete outcomes (binary or categorical dependent variables)
     Binary outcome: standard logistic regression
     i.e., Pass (1) or Fail (0)
     Categorical outcome: multinomial logistic regression
     i.e., Engineer (1) or Doctor (2) or Lawyer (3) or Teacher (4)

    In general, logistic regression relates to the way predictor variables impact


    the probability of any event and these significant variables identify the
    most promising target customer or most critical variable combinations.
    Binary Logistic Regression

     The logistic equation (cumulative logistic distribution (logit)) is written as a


    function of z, where z is a measure of the total contribution of each variable x used to
    predict the outcome

    Probability
    (event)

     Coefficients determined by maximum likelihood estimation (MLE), so larger sample


    sizes are needed than for OLS
    Graph of the Logistic Function

    Probability (Event)

    z
    Mathematical Interpretation of Logistic Regression

    Adopting S-shaped curve gives the probability of purchase p on the y-axis

    Where
    Mathematical Interpretation of Logistic Regression
    Mathematical Interpretation of Logistic Regression…
    Model Fit

     Smaller the -2LL -> Higher the likelihood

     Cox and Snell R-square


    Close to 1 as possible
     Nagel-kerke R-square
    Forward LR Method

     In Forward LR approach, the algorithm starts with a model having no


    independent variable, only a constant term. This is the Block 0 or Beginning
    Block.

     Therefore, one by one predictor variables are added, whichever gives most
    decrease in -2LL

     At every stage model Chi-square is computed which is the difference between


    -2LL for previous model and the current model. If this difference is not
    significant anymore, the algorithm stops.
    Forward LR Method

     Also, at any stage the probability of the event is computed


    Forward LR Method
    Example
    No. Loyalty Brand Product Shopping
    1 1 4 3 5
    2 1 6 4 4
     Brand_Loyalty_Data.xlsx 3
    4
    1
    1
    5
    7
    2
    5
    4
    5
    5 1 6 3 4
    6 1 3 4 5
    7 1 5 5 5
    8 1 5 4 2

     Output Variable: Loyalty


    9 1 7 5 4
    10 1 7 6 4
    11 1 6 7 2
    12 1 5 6 4
     Input Variables: Brand, Product, Shopping 13
    14
    1
    1
    7
    5
    3
    1
    3
    4
    15 0 7 5 5
    16 0 3 1 3
    17 0 4 6 2
    18 0 2 5 2
    19 0 5 2 4
    20 0 4 1 3
    21 0 3 3 4
    22 0 3 4 5
    23 0 3 6 3
    24 0 4 4 2
    25 0 6 3 6
    26 0 3 6 3
    27 0 4 3 2
    28 0 3 5 2
    29 0 5 5 3
    30 0 1 3 2
    Logistic Regression in R (1of3)

    library(readxl)
    library(MASS)
    mydata<-read_excel("Brand_Loyalty_Data.xlsx")
    #Print Data Frame
    mydata

    #Call glm function for Logistric Regression

    mymodel<-glm(Loyalty~ Brand+Product+Shopping, data = mydata)

    summary(mymodel)
    #mymodel$linear.predictors #Use this to see the z-values
    #mymodel$coefficients #Use this to see the coefficients
    Logistic Regression in R (2of3)

    Output

    Interpreting the Output:


    Z=-0.475180+0.161153*Brand+0.007827*Product+0.047897 *Shopping
    P=1/(1+e-z)
    Logistic Regression in R (3of3)

    Confusion matrix and Hit Ratio (Accuracy) of the model


    #Predicting the case/cases:
    #Predicting the case/cases:
    z_values<-predict(mymodel, mydata) #Here we consider training data. One can use test data
    #One can use mymodel$linear.predictors to get
    z_values directly
    z_values
    Loyalty_Predicted=1/(1+exp(-z_values))>0.5
    Output (Confusion Matrix)
    table(ActualValue=mydata$Loyalty, PredictedValue= Loyalty_Predicted)
    Hit Ratio (Accuracy)=Total Correct Prediction/(Total Correct Prediction
    + Total False Prediction)
    =16/(16+14)
    =53.3%
    Logistic Regression in R (3of3)

    Predicting a case
    #Predicting a case : Brand=1, Product=3, Shopping=2
    l1=list("Brand"=1, "Product"=3, "Shopping"=2) #One can use data frame instead of a list
    z<-predict(mymodel, newdata = l1)
    z
    prob=1/(1+exp(-z))
    prob Output
    PredictedValue=1/(1+exp(-z))>0.5
    PredictedValue

    Z value Probability Predicted value (Class)

    You might also like