Introduction to Machine Learning Concepts

Machine Learning Definitions

• Training set: A dataset of input-output pairs given to a machine learning model to create a prediction model or hypothesis.
• Hypothesis: A function learned from training data, which predicts an output based on an input.

Supervised, Unsupervised, and Reinforcement Learning

• Supervised learning: Uses labelled data (input-output pairs) to train a model to predict the output for new input data.
• Unsupervised learning: Uses unlabelled data to find patterns or clusters in the data. The model learns to identify structures without specific output guidance.
• Reinforcement learning: The learning system interacts with an environment and receives feedback, in the form of rewards or penalties, to learn the optimal behavior or strategy to maximize rewards.

Key Concepts

• Linear regression: A supervised learning algorithm used to predict a continuous target variable based on one or more independent variables.
• Logistic regression: A supervised learning algorithm used for classification tasks, predicting the probability of an input belonging to a particular category.
• Neural network: A type of machine learning model inspired by the structure of the human brain, consisting of interconnected nodes (neurons) organized in layers.
• Backpropagation: An algorithm used for training neural networks by adjusting weights to reduce error. It propagates the error backwards from the output layer to the input layer.
• Gradient descent: An optimization algorithm used to find the minimum of a function by iteratively moving in the direction of the negative gradient.
• Clustering: An unsupervised learning technique used to group data points into clusters based on their similarity based on their characteristics.
• Centroid: The center or average of a cluster, used as a representative point for the cluster in k-means clustering.

Bias and Variance

• Bias: A model's tendency to consistently under- or over-predict values. High bias implies the model is too simple and cannot capture the underlying relationship effectively.
• Variance: A model's sensitivity to changes in the training data. High variance implies the model is too complex, and may overfit the training data, leading to poor generalization to new data.

Key Terms

• RMSE (Root Mean Squared Error): A measure of the difference between the predicted values and the actual values, used to evaluate the performance of regression models.
• Correlation coefficient: A statistical measure that quantifies the strength and direction of the linear relationship between two variables.

Neural Network Concepts

• ReLU (Rectified Linear Unit): A type of activation function used in neural networks. It outputs the input directly if it is positive and zero if it is negative.
• Vanishing gradient problem: A common problem in deep neural networks with sigmoid activation functions, where gradients become very small as information propagates through the layers, slowing down training.

Cross-validation

• K-fold cross-validation: A technique used to evaluate the performance of a machine learning model by splitting the data into k folds and training the model on k-1 folds and validating on the remaining fold.
• Stochastic Gradient Descent (SGD): An optimization algorithm used for training machine learning models, which updates the weights using a single data point (or a small batch of data) at a time, instead of the entire dataset. This can be faster in practice compared to using the full dataset, but it often leads to higher variance in training.

Important Notes

• Overfitting occurs when a model performs well on the training data but poorly on unseen data.
• Increasing the complexity of a neural network by adding layers, increasing hidden layer size, or reducing regularization strength can lead to overfitting.
• K-means clustering is an unsupervised learning problem with a goal to find clusters of data points.