What is the main difference between machine learning and traditional programming?

1.  Machine learning does not need any data
2.  Traditional programming learns rules from examples
3.  In machine learning the rules are learned from data and example outputs instead of being written by humans
4. Machine learning always outperforms hand-written rules

A logistic regression model trained on two features has weights w₀ = −6, w₁ = 1, w₂ = 1.

Write the equation of the decision boundary and describe its shape.

Classify each test point below with threshold 0.5, showing z and p.

Try different values of K (e.g., 1, 3, 5, 10). How does changing K affect the accuracy? Change the test_size in the train_test_split to 0.3 or 0.1. Does a larger test set affect performance?

does crossvalidation will give you new  information for the k values?

https://colab.research.google.com/drive/1G9ng-Sj4NP8mNFjvrhY1owfxPyeUYjEv#scrollTo=rV-bdz13UBKh

What is cross-validation, and why is it important?

Let (1, -1) and (1, -5)  be samples from a class A and (3, 0), (1, 1) and (3, 2) samples from a class B.   By using the  KNN (K=5) classifier assign a class (A or B) to the sample (1,1).

Write a function that receive two vectors : a vector of the model parameters and a vector corresponding to a sample to be classified. The function returns a 1 or 0. Assume the classifier is a logistic regression classifier. 

The

following model was trained to predict whether a fruit is a pear (0) or apple

(1). The training  samples contain information about the diameter and the two axes ratio for each sample.

All classifiers are described by linear models

Linear regression is a Deep Learning model

A scientist is seeking to calculate the probability of rain for each day in November 2024 in Puerto Rico. Historical data for the same month over the past 10 years is available. Which algorithm would you recommend for this analysis?