lOMoARcPSD|28201683
Ml-2 - Notes
Machine learning (Visvesvaraya Technological University)
Scan to open on Studocu
Studocu is not sponsored or endorsed by any college or university
Downloaded by DIVYA SHREE (divyashree@[Link])
lOMoARcPSD|28201683
Downloaded by DIVYA SHREE (divyashree@[Link])
lOMoARcPSD|28201683
Understanding Data and Basics of Learning Theory
6th
BCS602
Module -2
Downloaded by DIVYA SHREE (divyashree@[Link])
lOMoARcPSD|28201683
1. Explain the process of obtaining principle components and its relevance in feature
reduction..
Principal Component Analysis (PCA)
✔ The Principal Component Analysis (PCA) or KL Transform is used to transform a given set of
measurements into a new set of features.
✔ These features exhibit high information packing properties, leading to a reduced and compact feature
set.
✔ The elimination of redundant information allows for a lower-dimensional representation of data
Process of Obtaining Principal Components
1. Representation of Data as Random Vectors
Consider a group of random vectors of the form:
[Link] support@[Link]
Downloaded by DIVYA SHREE (divyashree@[Link])
lOMoARcPSD|28201683
2. Compute the Mean Vector
3. Compute the Covariance Matrix
[Link] support@[Link]
Downloaded by DIVYA SHREE (divyashree@[Link])
lOMoARcPSD|28201683
4. Approximate Mean and Covariance Matrix for Large Datasets
5. Compute Eigenvalues and Eigenvectors
[Link] support@[Link]
Downloaded by DIVYA SHREE (divyashree@[Link])
lOMoARcPSD|28201683
6. Construct the Transform Kernel
The transform kernel (A) is built using the eigenvectors of the covariance matrix.
7. Transform the Data
8. Reconstruction of the Original Data
[Link] support@[Link]
Downloaded by DIVYA SHREE (divyashree@[Link])
lOMoARcPSD|28201683
9. Feature Reduction and Selection of Principal Components
[Link] support@[Link]
Downloaded by DIVYA SHREE (divyashree@[Link])
lOMoARcPSD|28201683
2. Explain the procedure for pair-t tests and Chi-Square goodness fit test.
[Link] support@[Link]
Downloaded by DIVYA SHREE (divyashree@[Link])
lOMoARcPSD|28201683
[Link] support@[Link]
Downloaded by DIVYA SHREE (divyashree@[Link])
lOMoARcPSD|28201683
3.
[Link] support@[Link]
Downloaded by DIVYA SHREE (divyashree@[Link])
lOMoARcPSD|28201683
4.
[Link] support@[Link]
Downloaded by DIVYA SHREE (divyashree@[Link])
lOMoARcPSD|28201683
5. Explain the procedure for hypothesis testing?
✔ Data collection alone is not enough. Data must be interpreted to give a conclusion. The
conclusion should be a structured outcome.
✔ This assumption of the outcome is called a hypothesis.
✔ Statistical methods are used to confirm or reject the hypothesis. The assumption of the
statistical test is called the null hypothesis. It is also called as hypothesis zero (H0).
✔ In other words, hypothesis is the existing belief. The violation of this hypothesis is
called first hypothesis (H1) or hypothesis one.
There are two types of hypothesis tests, parametric and non-parametric.
✔ Parametric tests are based on parameters such as mean and standard deviation.
✔ Non-parametric tests are dependent on characteristics such as independence of events or
data
There are two types of Error involved in statistical test
✔ Type I error is the incorrect rejection of a true null hypothesis and is called false positive.
✔ Type II error is the incomplete failure of rejecting a false hypothesis and is called false
negative.
[Link] support@[Link]
Downloaded by DIVYA SHREE (divyashree@[Link])
lOMoARcPSD|28201683
[Link] support@[Link]
Downloaded by DIVYA SHREE (divyashree@[Link])
lOMoARcPSD|28201683
p-value
✔ A probability value used to decide acceptance or rejection of a hypothesis.
✔ Determines the statistical significance of a test.
✔ Example:
[Link] p-value = 0.03 and significance level (α) = 0.05, then 0.03 < 0.05, so the result is statistically significant.
✔ Decision rule:
[Link] p-value ≤ α, reject H0 (null hypothesis).
[Link] p-value > α, reject H1 (alternative hypothesis).
✔ Significance level (α): Standard threshold values include 0.05, 0.01, etc.
[Link] support@[Link]
Downloaded by DIVYA SHREE (divyashree@[Link])
lOMoARcPSD|28201683
6.
[Link] support@[Link]
Downloaded by DIVYA SHREE (divyashree@[Link])
lOMoARcPSD|28201683
7.
[Link] support@[Link]
Downloaded by DIVYA SHREE (divyashree@[Link])
lOMoARcPSD|28201683
[Link] support@[Link]
Downloaded by DIVYA SHREE (divyashree@[Link])
lOMoARcPSD|28201683
8. Explain the key steps involved in the design of a learning system. Why is each
step important in building an effective learning algorithm?
The design of a learning system involves four key steps. Each step is crucial for developing an efficient
learning algorithm. Let's go through them in detail:
1. Choosing a Training Experience
✔ This step involves selecting the data or experience from which the system will learn.
✔ The training experience should be representative of the real-world problem the system is trying to
solve.
✔ Example: In a spam detection system, training data would include emails labeled as "spam" or "not
spam.“
2. Choosing a Target Function
✔ The target function defines the goal of learning, i.e., what the system is trying to predict or classify.
✔ It maps inputs (features) to outputs (labels or decisions).
[Link] support@[Link]
Downloaded by DIVYA SHREE (divyashree@[Link])
lOMoARcPSD|28201683
[Link] support@[Link]
Downloaded by DIVYA SHREE (divyashree@[Link])
lOMoARcPSD|28201683
9. Explain hypothesis space search by Find –S Algorithm and Limitations.
[Link] support@[Link]
Downloaded by DIVYA SHREE (divyashree@[Link])
lOMoARcPSD|28201683
Limitations of Find-S Algorithm
1. Ignores Negative Instances
[Link] Find-S algorithm finds a hypothesis that is consistent with only positive instances and ignores all
negative instances.
[Link] the training dataset is consistent, the hypothesis found may also be consistent.
2. Finds Only One Unique Hypothesis
[Link] may be multiple hypotheses that fit the training data, but Find-S outputs only one specific
hypothesis, potentially missing other valid solutions.
3. Sensitive to Errors in Data
[Link] the training dataset contains errors, Find-S can be misled, as it does not account for negative examples.
Solution: The Candidate Elimination Algorithm was introduced to address these issues by considering
both positive and negative examples, generating a hypothesis set instead of a single hypothesis.
[Link] support@[Link]
Downloaded by DIVYA SHREE (divyashree@[Link])
lOMoARcPSD|28201683
10. Consider the training dataset of 4 instances shown in Table 3.2. It contains the details of the performance of students
and their likelihood of getting a job offer or not in their final semester. Apply the Find-S algorithm.
[Link] support@[Link]
Downloaded by DIVYA SHREE (divyashree@[Link])
lOMoARcPSD|28201683
[Link] support@[Link]
Downloaded by DIVYA SHREE (divyashree@[Link])
lOMoARcPSD|28201683
Solve this problem using find S-Algorithm
Solution : <sunny, warm, ?, strong, ?, ?>
Downloaded by DIVYA SHREE (divyashree@[Link])
lOMoARcPSD|28201683
11. Difference between Bias and Variance
[Link] support@[Link]
Downloaded by DIVYA SHREE (divyashree@[Link])
lOMoARcPSD|28201683
12. What are the three types of prediction errors in machine learning, and how do
they impact model performance?.
1. Irreducible Error
• Cannot be reduced or avoided.
• Caused by unknown variables, noise, etc.
• Sometimes mitigated by data cleaning.
2. Bias Error
• Difference between predicted and actual values.
• Indicates accuracy of model predictions (related to Mean Square Error).
• Caused by inaccurate assumptions or underfitting.
• More prominent when testing the model with new data.
3. Variance Error
• Sensitivity of the model to noise, outliers, or small fluctuations in training data.
• Arises due to changes in parameters and training sets.
• Leads to variation in model predictions across different datasets.
[Link] support@[Link]
Downloaded by DIVYA SHREE (divyashree@[Link])
lOMoARcPSD|28201683
Join LearnyHive today and embark on a rewarding
journey of growth and academic achievement.
[Link]
LearnyHive
Hive Your Way to Educational Excellence.
/learnyhive @learnyhive
/company/learnyhive
Don't miss out on this opportunity!
Follow Us on Instagram and YouTube for regular updates.
Downloaded by DIVYA SHREE (divyashree@[Link])