Scribd
Upload
45 views4 pages

Class Notes: The Basics of Machine Learning

Uploaded by

Hardana Gading Perdana Putra
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
45 views4 pages

Class Notes: The Basics of Machine Learning

Uploaded by

Hardana Gading Perdana Putra
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
You are on page 1/ 4

Class Notes: The Basics of Machine Learning

1. Introduction to Machine Learning (ML)

 Definition: Machine Learning is a branch of artificial intelligence (AI) that enables


computers to learn from data and improve their performance without being explicitly
programmed.
 Goal: The aim is to develop algorithms that can recognize patterns, make decisions,
and predict outcomes based on data.

2. Types of Machine Learning

 Supervised Learning:
o Involves labeled data, where the model learns from input-output pairs.
o Common tasks include classification (e.g., spam detection) and regression
(e.g., predicting house prices).
 Unsupervised Learning:
o Uses unlabeled data, allowing the model to find hidden patterns or groupings
without specific output labels.
o Common tasks include clustering (e.g., customer segmentation) and
association (e.g., market basket analysis).
 Reinforcement Learning:
o Involves an agent learning to make a sequence of decisions to maximize a
reward.
o Common in robotics and game AI (e.g., autonomous driving).

3. Key Concepts in Machine Learning

 Training Data:
o Data used to teach the model. It contains examples the model will learn from.
 Testing Data:
o Data used to evaluate the model's performance after training.
 Features:
o The input variables used to make predictions. For instance, in a housing price
model, features might include square footage, location, and number of
bedrooms.
 Labels:
o The outputs associated with the training data, only present in supervised
learning.

4. Common Machine Learning Algorithms


 Linear Regression:
o Used for predicting continuous values by finding the linear relationship
between input features and output labels.
 Logistic Regression:
o Used for binary classification (e.g., yes/no, spam/not spam).
 Decision Trees:
o Use a tree-like model of decisions, helpful in classification and regression
tasks.
 K-Nearest Neighbors (KNN):
o A simple, instance-based learning algorithm that classifies data based on the
'k' closest data points.
 Support Vector Machines (SVM):
o Creates a hyperplane in a high-dimensional space to separate classes of data.
 Neural Networks:
o Consist of interconnected layers of nodes, inspired by the human brain,
commonly used in complex tasks such as image and language processing.

5. The Machine Learning Process

 Data Collection:
o Gather relevant data to solve the problem.
 Data Preprocessing:
o Clean the data, handle missing values, and scale features for better
performance.
 Feature Engineering:
o Selecting or creating features that improve model accuracy.
 Model Selection:
o Choosing the right algorithm for the task (e.g., linear regression, neural
networks).
 Training:
o The process of feeding data to the model and adjusting its parameters.
 Evaluation:
o Testing the model’s performance using metrics like accuracy, precision, recall,
and F1 score.
 Optimization:
o Fine-tuning model parameters to improve performance.

6. Evaluation Metrics

 Accuracy:
o The percentage of correct predictions in the total predictions.
 Precision:
o The ratio of true positive predictions to the total positive predictions.
 Recall:
o The ratio of true positive predictions to the total actual positives.
 F1 Score:
o A balance between precision and recall, especially useful for imbalanced
datasets.

7. Overfitting and Underfitting

 Overfitting:
o Occurs when the model learns the noise in the training data, performing well
on training data but poorly on new data.
o Solutions include using regularization, simplifying the model, or collecting
more data.
 Underfitting:
o Occurs when the model is too simple to capture patterns in the data.
o Can be improved by using a more complex model or including additional
features.

8. Applications of Machine Learning

 Natural Language Processing (NLP):


o Powering applications like chatbots, language translation, and sentiment
analysis.
 Computer Vision:
o Used for facial recognition, object detection, and medical imaging.
 Healthcare:
o Assisting in diagnosis, drug discovery, and personalized medicine.
 Finance:
o Fraud detection, stock price predictions, and credit scoring.
 E-commerce:
o Recommender systems that suggest products to users based on previous
behavior.

9. Challenges in Machine Learning

 Data Quality:
o Good data is critical, and issues like missing or noisy data can impact
performance.
 Bias and Fairness:
o Models may inherit biases present in the training data, leading to unfair
predictions.
 Computational Costs:
o Training large models, especially deep learning models, requires significant
resources.
 Interpretability:
o Complex models (e.g., neural networks) are often hard to interpret, which is
crucial for certain applications like healthcare.
Summary: Machine learning is transforming multiple fields by enabling data-driven
decision-making and automation. Its success depends on high-quality data, careful selection
of algorithms, and a structured approach to model evaluation and improvement.

You might also like