DATA MINING GLOSSARY

Find out meanings of the most important data mining terms.

Antecedent	Antecedent is a part of association. When association is determined between two events, the first event is called antecedent. For example "If customer buys pizza he also buys bear or lemonade in 89% cases", here "buys pizza" is the antecedent.
Association	An association algorithms aim is to create decision rules that describe how often events have occurred together. For example, "When customers purchase a PC, they also purchase a monitor 87% of the time." These relationships are mostly created with a confidence interval.
Backpropagation	Backpropagation is a neural network training method. The neural network is adjusted until the difference between actual output and desired output is minimal.
Classification	Classification is a process of determination of a predefined class, to which an example belongs. For example, given classes of clients that correspond to loyal and reliable ones, and the others, identify from a data on a new client to which class she/he belongs.
Classification tree	Classification trees place examples and categorial variables  into pre-defined classes.
Data	Data is information, facts, figures and values. Data can be collected by observation, different measurements, experiments, surveys, etc.
Data Mining	An information extraction activity. Data mining  goal is to discover new knowledge, revealing  hidden facts contained in databases. Some Data mining methods are: statistical analysis, neural networks, machine learning, modelling, database technologies. The result of Data mining processing are often the decision rules, used to predict future results.
Data Warehouse	A data warehouse is a collection of data, structured and designed for further querying, decision making, data mining and knowledge discovery.
Decision Trees	A tree-like way of representing a collection of hierarchical rules, that are nodes of the Tree, where a decision has to be made. Final nodes represent classes or values.
Deduction	Deduction concludes information that is a logical consequence of the processed data.
Discrete Data	Discrete data consists of a finite set of values.
Leaf Node	Is a node, that cannot be further splitted. It represents a class, a group or a decision in Decision Trees.
Node	A node in a Decision or Classification Tree. A node represents a decision, grouping or condition for further splitting.
Pattern Recognition	Pattern recognition is a process o data classification. The process is usually based on statistical information, extracted from data. Pattern Recognition aim is to detect relationships between variables. Data mining techniques use automatic pattern discovery, detecting complicated non-linear relationships in data. The patterns are usually extracted from measurements and observations data.
Prediction	Prediction (also called regression) is similar to classification. The only difference is that in prediction the target attribute is not discrete but a continuous one. Prediction is processed to discover the numerical value of the target attribute for selected examples.
Predictive Models	Examples of predictive models are decision trees, neural networks, logistic regression or rule based models. The predictive model is used to obtain scoring. The characteristics are inputted as a vector into the model, and as a result we receive an output score.
Training data	This data is used to create or train a model.
Validation	Validation is testing already obtained decision rules, decision trees or other models on a data, that differs from the training data.
Validation Set	A validation set is a part of data, used for data mining. The data set is usually divided into two parts: a training (also called learning) data set and validation data set. While the training data set is used for building a model, the validation set is used to test models efficacy.
Variance	Variance is a measure of dispersion from average values.
Visualisation	Visualisation is a process of graphical displaying data. Data can be represented as graphs, charts, models, etc.