BMAN73701 Week 5: Advanced Machine Learning

The ratio of correctly predicted positive instances to the total number of instances predicted as positive.

The ratio of correctly predicted positive instances to the total number of actual positive instances.

The ratio of correctly classified instances to the total number of instances.

A measure that combines precision and recall to give a balanced evaluation of a model.

A table that summarizes the performance of a classification model by showing the counts of true positives, true negatives, false positives, and false negatives.

An instance that is correctly predicted as positive.

An instance that is correctly predicted as negative.

An instance that is incorrectly predicted as positive.

Data in its original, unprocessed form. It can be in various formats like databases, spreadsheets, text files, or web data.

The process of cleaning and organizing raw data to prepare it for analysis. This involves dealing with inconsistencies, missing values, and transforming data into a usable format.

Data organized in a table format, with rows representing observations and columns representing variables.

Data that represents quantities or measurements, like age, weight, or income.

Data that represents categories or groups, like gender, color, or product type.

A type of categorical data where the categories have a natural order or ranking, like small, medium, large.

Transforming and preparing data for machine learning algorithms. This includes scaling, encoding, and handling missing values.

The process of examining and interpreting data to gain insights and make informed decisions.

A type of machine learning where the algorithm learns from labeled data, meaning it's given both inputs and expected outputs. It aims to predict outputs based on the provided examples.

A type of machine learning where the algorithm learns from unlabeled data, meaning it has to discover patterns and structures on its own. It aims to find insights and relationships within the data.

A type of supervised learning that categorizes data into distinct groups or classes.

A type of supervised learning that predicts a continuous value (like a number) based on given input data.

The process of dividing the dataset into two parts: one for training the machine learning model and another for evaluating its performance.

The part of the dataset used to evaluate the performance of the trained machine learning model.

A metric used to assess the performance of a machine learning model on the test data.

A robust technique for evaluating a model by repeatedly splitting the training data into K folds, training the model on K-1 folds, and testing on the remaining fold.

A portion of the training data used to evaluate the performance of the model during training, helping to prevent overfitting.

The first layer of a neural network that receives the raw input data.

A processing layer in a neural network that transforms the input data and learns complex features.

The final layer of a neural network that produces the predicted output.

A complex machine learning model inspired by the structure of the human brain, composed of interconnected nodes called neurons.

A machine learning algorithm that combines multiple decision trees to improve accuracy and prevent overfitting.

A flowchart-like structure that uses a series of rules to classify data points based on their features.

A group of multiple learning models working together to improve prediction accuracy.

In Random Forests, each tree is built independently, making random decisions about which features to use and how to split the data.

After each tree in the Random Forest makes a prediction, these predictions are combined through voting (for classification) or averaging (for regression) to arrive at the final output.

A measure used in decision tree construction to determine the best feature to split the data based on how much it improves the separation of classes.

In regression, the variance of data points around the prediction line is used to evaluate the accuracy of the prediction.

A scenario where a model performs well on the training data but poorly on unseen data due to learning the noise and specific patterns of the training data.

A type of machine learning where an algorithm learns from labeled data, meaning each data point has a known output or target value. This allows the model to learn the relationship between inputs and outputs, making predictions on new, unseen data.

The process of feeding labeled data to a machine learning model to adjust its parameters and improve its ability to make accurate predictions.

The process of finding the best set of parameters (hyperparameters) for a machine learning model to achieve optimal performance.

A sequence of steps used to process data and train a machine learning model. It typically involves data cleaning, preprocessing, feature engineering, model selection, and evaluation.

If K is too small, there's less time but poorer generalization. Training might not have enough examples, leading to inaccurate performance predictions on unseen data.

When K is too large, it takes longer and increases variance. The validation set might be too small, leading to unstable performance estimates.

A technique used for classification problems, where the proportion of classes is maintained across train and test sets, as well as in each fold. Ensures equal class representation for better model evaluation.

When dealing with unbalanced classes (more samples from one group), stratification ensures the model doesn't favor the majority class and leads to more reliable results.

This parameter in train_test_split() tells the function to split the data into train and test sets, maintaining the proportion of class labels in each set.

This function automatically utilizes stratified k-fold cross-validation to assess models, ensuring balanced class representation across folds.

K-fold cross-validation allows for robust performance evaluation of a model by assessing its generalization ability across different folds, ensuring it performs well on unseen data.