Machine Learning
A subfield of Al involving algorithms that iteratively learn from and then make decisions, recommendations, inferences or predictions based on input data. These algorithms build a model from training data to perform a specific task on new data without being explicitly programmed to do so. Machine learning (ML) implements various algorithms that learn and improve by experience in a problem-solving process that includes data collection and preparation, feature engineering, training, testing and validation. Machine learning detects patterns in data to learn tasks. It improves performance through experience.
Key ML approaches include:
Deep Learning: This is a subset of ML that utilizes multilayered neural networks to learn from large amounts of data, as might be found in radiologic images or genomic sequences.
Supervised Learning: In this learning style, algorithms learn from a training dataset where the correct output (label) is already known.
Unsupervised Learning: In this learning style, algorithms detect hidden patterns in data where the output is unknown. This is highly useful for identifying patient clusters.
Classification: This is a prediction technique where models predict a category. An example is predicting if a patient is "high risk" versus "low risk" for readmission.
Regression: This is a prediction technique where models predict a continuous numerical value. An example is predicting a patient's specific length of stay.
Validation: This refers to the necessity of using separate training, validation, and test datasets. This practice ensures a model is accurate and generalizable to new patients.