L2 - Types Of Learning Flashcards
(15 cards)
Supervised learning can be used to map from input data to an output, what is the difference between Classification and Regression in this instance ?
In Classification the output is discrete values (e.g., “spam” or “not spam”, “cat” or “dog”).
In regression the goal is to predict a continuous numerical value. The output is a set of real numbers (e.g., price, temperature, age).
What is Supervised Learning?
Training algorithms using labeled datasets, where each input is paired with the correct output. This allows the algorithm to learn the mapping from inputs to outputs, enabling it to make predictions or decisions when presented with new, unseen data.
Examples:
* Classification
* Regression
What is unsupervised learning?
Training algorithms using labeled datasets, unsupervised learning algorithms are tasked with finding patterns and relationships within the data without any prior knowledge of the data’s meaning.
Examples:
* Clustering
* Density Estimation
Give three examples of classification algorithms
- Spam Filtering
- Object Detection
- Weather Prediction
What is Clustering?
Grouping Data into clusters depending on similarity
What is Classification?
Assigning classes to Data
What is Regression?
mapping input to output by a set of real numbers.
Eg The relation between hand and eye data.
Give examples of Regression algorithm use
- Stock market trend and prediction
- Machinery lifetime estimate
Name the four classification methods covered by the module.
- Support Vector Machines
- Discriminant Analysis
- Naive Bayes
- Nearest Neighbour
name the five Regression methods covered by the module.
- Linear Regression GLM
- SVR GPR
- Ensemble methods
**4. Decision trees - Neural Networks**
Name the five Clustering methods covered by the course.
- K-means, Kmedoids, Fuzzy C-means
- Hierarchical
- Gaussian mixture
- Neural networks
- Hidden markov model
What is K-means clustering?
A method of dividing observations into a given number of groups by determining the location of k group centers.
Then assigning each observation to a group according to which center the it is closest to.
How do you determin k group centers in k-means clustering?
- Start with random centre locations
- Assign the data to the closest centres
- For each group, calculate the groups actual average centre and update the k centre to this location
- Repeat until the centres are fixed
5.(optional) repeat for new starting centres. For each solution calculate the total distance between data and their closest centre. The optimal solution will have the lowest total distance.
Hierarchical Clustering
Hierarchical clustering is a method of cluster analysis that seeks to build a hierarchy of clusters. It’s commonly used in data mining and statistics to group similar objects into clusters based on their characteristics.
Gaussian Mixture Models
A Gaussian Mixture Model (GMM) is a probabilistic model used to represent a dataset as a mixture of several Gaussian (normal) distributions.
Instead of assigning each data point to a single cluster (like K-Means), GMM assumes that:
The data is generated from a mixture of several Gaussian distributions.
Each data point has a probability of belonging to each cluster.
