- [Colorado State - Forbes]
- Overview
A machine learning (ML) model is a program that finds patterns or makes decisions from previously unseen data sets.
For example, in natural language processing, ML models can parse and correctly identify the intent behind previously unheard sentences or word combinations. In image recognition, a ML model can be trained to recognize objects -- such as cars or dogs. ML models can perform such tasks by being "trained" with large datasets.
During training, ML algorithms are task-optimized to find specific patterns or outputs from data sets. The output of this process - usually a computer program with specific rules and data structures - is called a ML model.
How many models are there in ML? Many! ML is an evolving field and there are always more ML models being developed.
- Machine Learning Algorithms
A ML algorithm is a mathematical method of finding patterns in a set of data. ML algorithms typically come from statistics, calculus, and linear algebra. Some popular examples of ML algorithms include linear regression, decision trees, random forests, and XGBoost.
- Model Training in Machine Learning
The process of running a machine learning algorithm on a data set (called training data) and optimizing the algorithm to find a specific pattern or output is called model training. The resulting function with rules and data structure is called a trained machine learning model.
- Different Types of Machine Learning
In general, most machine learning techniques can be categorized as supervised learning, unsupervised learning, and reinforcement learning
- Decision Trees in Machine Learning (ML)
A decision tree is a predictive method in ML to determine which class an object belongs to. As the name suggests, a decision tree is a tree-like flow diagram in which the class of an object is determined step-by-step using certain known criteria.
- Regression in Machine Learning
Regression in data science and machine learning is a statistical method that predicts an outcome based on a set of input variables. The result is usually a variable that depends on a combination of input variables.
- Classifiers in machine learning
A classifier is a machine learning algorithm that assigns objects as members of a class or group. For example, classifiers are used to detect whether an email is spam, or whether a transaction is fraudulent.
- The Best Model for Machine Learning
The best machine learning model for a particular situation depends on the desired outcome. For example, to predict the number of vehicle purchases in a city based on historical data, supervised learning techniques such as linear regression may be most useful. On the other hand, to determine whether a potential customer in that city would purchase a vehicle, a decision tree might be most effective given their income and commuting history.
- Model Deployment in Machine Learning (ML)
Model deployment is the process of making a machine learning model available to the target environment (for testing or production). The model is often integrated with other applications in the environment (such as databases and UIs) through APIs. Deployment is the stage where an organization can really reap the rewards of a large investment in model development.
- Deep Learning Models
Deep learning models are a class of ML models that mimic the way humans process information. The model consists of several layers of processing (hence the name "deep") to extract high-level features from the data provided. Each processing layer passes a more abstract representation of the data to the next layer, with the last layer providing more human-like insights. Unlike traditional ML models, which require labeled data, deep learning models can ingest large amounts of unstructured data. They are used to perform more human-like functions, such as facial recognition and natural language processing.
- Time Series Machine Learning
A time series machine learning model is a model where one independent variable is a continuous length of time (minutes, days, years, etc.) and has an effect on the dependent or predictor variable. Time series machine learning models are used to predict time-bound events such as - weather in the next week, expected number of customers in the next month, revenue guidance for the next year, etc.
