0% found this document useful (0 votes)
6 views82 pages

Machine Learning Algorithm Evaluation Techniques

The document outlines guidelines for conducting machine learning experiments, focusing on cross-validation techniques, performance metrics, and statistical tests such as the t-test and McNemar's test. It provides a step-by-step approach to performing an independent t-test to compare the performance of two models, including hypothesis formulation and data interpretation. Additionally, it discusses the application of McNemar's test for evaluating classifiers on paired nominal data.

Uploaded by

scarlsuresh
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
6 views82 pages

Machine Learning Algorithm Evaluation Techniques

The document outlines guidelines for conducting machine learning experiments, focusing on cross-validation techniques, performance metrics, and statistical tests such as the t-test and McNemar's test. It provides a step-by-step approach to performing an independent t-test to compare the performance of two models, including hypothesis formulation and data interpretation. Additionally, it discusses the application of McNemar's test for evaluating classifiers on paired nominal data.

Uploaded by

scarlsuresh
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd

UNIT-2

DESIGN AND ANALYSIS OF


MACHINE LEARNING
ALGORITHMS
Contents
• Guidelines for machine learning experiments,
• Cross Validation (CV) and resampling –
• K-fold CV, bootstrapping, measuring classifier performance, assessing a
single classification algorithm and comparing two classification
algorithms
• t test, McNemar’s test, K-fold CV paired t test Performance
metrics-MSE, accuracy, confusion matrix, precision, recall, F1- Score
• Linear Regression with multiple variables-
• Logistic Regression-
• spam filtering with logistic regression
t-test
• The t-test is a statistical hypothesis test used to determine whether

there is a significant difference between the means of two groups.

• In the context of machine learning, the t-test can be used to compare

the performance of two different models, two algorithms, or the same

model under different conditions.


• Types of t-Tests

1. Independent t-test (Two-sample t-test): Used when comparing the means

of two independent groups.

2. Paired t-test: Used when comparing the means of the same group under two

different conditions.
• Step-by-Step Explanation of the Independent t-Test
• Let's go through the steps of performing an independent t-test with an
example.
• Scenario
• Assume we have two models, Model A and Model B, and we want to
compare their performances based on accuracy scores obtained from
10-fold cross-validation
• Step 1: Formulate Hypotheses

• Null Hypothesis (H0): There is no significant difference between the


performance of Model A and Model B.

• Alternative Hypothesis (H1): There is a significant difference


between the performance of Model A and Model B.
• Step 2: Collect Data
• Suppose we have the following accuracy scores for each fold:
• Model A: [0.85, 0.87, 0.88, 0.86, 0.89, 0.87, 0.88, 0.86, 0.89, 0.88]
• Model B: [0.83, 0.82, 0.84, 0.85, 0.84, 0.83, 0.85, 0.82, 0.83, 0.84]
• Step 3: Calculate Means and Standard Deviations
• Calculate the mean and standard deviation for both sets of scores.
• Step 4: Perform the Independent t-TestUse the ttest_ind function from
the [Link] library to perform the t-test.
• Step 5: Interpret the Results
• t-statistic: The t-statistic measures the size of the difference relative to
the variation in the sample data.
• p-value: The p-value indicates the probability of observing the results
given that the null hypothesis is true.
• Let's assume the following results were obtained:
• t_statistic, p_value = 5.57, 0.0001
• Interpretation:
• If the p-value is less than the chosen significance level (typically
0.05), we reject the null hypothesis. In this case, since the p-value is
0.0001, which is much less than 0.05, we reject the null hypothesis.
• This means there is a significant difference between the performance
of Model A and Model B.
McNemar’s Test: Explanation and
Example
• McNemar’s test is a statistical test used on paired nominal data to
determine whether there are differences in the proportions of two
related groups.

• It's commonly used in machine learning to compare the performance


of two classifiers on the same dataset, especially when the data is in
the form of a contingency table.
When to Use McNemar’s Test?
• Binary classification problems: When you have two classifiers and

want to compare their predictions.

• Paired samples: The same instances are classified by both classifiers,

so their predictions can be directly compared.

You might also like