Machine Learning | Supervised Learning vs Unsupervised Learning (Questions and Answers) #eduvictors #ipmusings
Machine Learning | Supervised Learning vs Unsupervised Learning (Questions and Answers)
Q: What is the difference between supervised and unsupervised machine learning?
Answer: Supervised learning requires training labelled data. For example, to do classification (a supervised learning task), you’ll need to first label the data you’ll use to train the model to classify data into your labelled groups. Unsupervised learning, in contrast, does not require labelling data explicitly.
The popular algorithms in supervised learning are categorized as either classification or regression models. In unsupervised learning, there is a clustering model, dimensionality reduction, and autoencoders.
Q: What are the types of predictive models that come under Supervised Learning?
Answer:
- Regression
- Classifier
Q: Give types of Regression Models.
Answer: Linear, Ridge & Lasso, Polynomial Regression, Multi-linear, Non-Linear, Time series forecasting.
Q: What is meant by labelled data?
Answer: Labeled data means the input data and its associated output is available in the dataset.
Q: What are the salient features of Supervised learning?
Answer: In supervised learning, an important method in machine learning, the models can be trained to make correct predictions depending on data containing the required labels.
Here are the salient features of supervised learning:
1. Labeled Data:
Supervised learning depends on labelled data, where each input is paired with the correct output.
Example: In a dataset of emails, each email is labelled as 'spam' or 'not spam'.
2. Training Process:
The algorithm is trained using the labelled data to learn the mapping between inputs and outputs.
Example: Training a model to recognise handwriting by providing images of handwritten digits and their corresponding numbers.
3. Prediction:
Once trained, the model can predict the output for new, unseen inputs based on the learned patterns.
Example: Using a trained model to predict house prices based on features like size, location, and number of bedrooms.
4. Feedback Loop:
Supervised learning involves a feedback loop where the model’s predictions are compared to the actual outcomes, and the model is adjusted accordingly. A beneficial feedback loop typically involves bringing in unbiased, external information into the machine learning system. Feedback loops can be either negative or positive.
Example: Backpropagation algorithms are used to identify inconsistencies and feed corrected information back into the model as input. This constant adjustment through feedback loops enables the ML model to learn from its mistakes and improve its accuracy.
5. Performance Metrics:
The performance of supervised learning models is evaluated using metrics such as accuracy, precision, recall, and F1 score.
Example: The F1 score is a machine learning (ML) metric for evaluating model accuracy, combining precision and recall.
6. Applications:
Supervised learning is widely used in various applications, including image recognition, speech recognition, and medical diagnosis.
Example: Using supervised learning to diagnose diseases from medical images.
7. Overfitting and Underfitting:
Overfitting occurs when the model learns the training data too well, including noise, and performs poorly on new data. Underfitting occurs when the model is too simple to capture the underlying patterns in the data.
Example: A student memorises maths problems from a book but fails to solve a problem with a similar pattern but having different values. (Overfitting). Another example of overfitting is a simplistic model that determines creditworthiness based simply on a person’s income without taking into account other factors such as their spending habits, prior loans and job history may lead to making poor lending decisions.
In underfitting, a student memorises summaries of the chapters but misses the important details.
8. Algorithms:
Common algorithms used in supervised learning include linear regression, logistic regression, decision trees, support vector machines, and neural networks.
Example: Using linear regression to predict a continuous outcome, like a student’s test score based on study hours.
DO YOU KNOW!
The term "machine learning" was coined by Arthur Samuel in 1959, an American pioneer in computer gaming and artificial intelligence. He stated that learning allows computers to improve without being explicitly programmed. Arthur Samuel created the first self-learning program for playing checkers, and you observe that the more the system plays, the better it performs.
Q. What is Unsupervised Learning?
Answer: Unsupervised learning is a type of machine learning where the algorithm learns from unlabeled data. Unlike supervised learning, where data is labelled and the algorithm learns to predict a target variable, unsupervised learning finds patterns and structures within the data without any prior guidance. This means that the data which is fed to the machine is random and there is no knowhow available about it to the trainer.
These learning models are used to identify trends, patterns, and relationships in the data that is fed into them. In this model, the major features are identified by the machine, which helps the user understand the data. For example, in a dataset of 100 cat images, if you want to discern some patterns, you would need to input this data into an unsupervised learning model and train the machine. Once trained, the machine would identify patterns in the data. These patterns might already be known to the user, such as colour or size, or they could reveal something unusual about the cats.
Key characteristics of unsupervised learning:
* Unlabeled data: No predefined categories or outputs.
* Pattern discovery: The algorithm identifies hidden patterns, relationships, and structures.
* Exploratory data analysis: Often used to explore and understand data.
Common techniques used in unsupervised learning are:
1. Clustering: Grouping similar data points together.
2. Association rule learning: Finding relationships between items.
3. Dimensionality reduction: Simplifying data by reducing the number of features.
4. Anomaly detection: Identifying unusual data points.
Applications of Unsupervised Learning:
1. Customer segmentation
2. Image and pattern recognition
3. Fraud detection
4. Market basket analysis
In essence, unsupervised learning is like giving a child a box of toys and asking them to find similarities and differences without any instructions. The child learns through exploration and observation, much like the algorithm in unsupervised learning.
👉See Also:
