Table Of Contents
  • Clustering
  • Pattern-based Classifications
  • Visualization using WEKA
  • Sequential Pattern Mining Method


A cluster can be defined as a group of items that are in the same class. Meaning, we group similar objects in one cluster and dissimilar objects in another cluster. In data mining, clustering involves grouping abstract objects into classes of similar objects. Clustering is preferred to classifications because it is more adaptable to changes. It can also be used to single out essential features that distinguish different groups. Clustering methods can be classified as grid-based, partitioning, hierarchical, density-based, model-based, and constraint-based. These methods are applied in several fields including market research, data analysis, pattern recognition, and many more.

Pattern-based Classifications

The use of patterns in data mining help researchers obtain models for domains that are structured. Pattern-based classifications provide more accurate and interpretable models. A frequent pattern-based classification describes the essence of data mining. It is a process that involves learning patterns of subsets emerging within a dataset and across instances. Pattern-based classification methods can be categorized under two dimensions. First, if the method executes pattern mining algorithms iteratively or it post-processes a pre-computed set of patterns. Secondly, whether the method depends on a model to select a pattern or independently selects a pattern.

Visualization using WEKA

WEKA is a popular data mining tool that can be used to perform several tasks and experiment with new methods over datasets. Data visualization is a method that is used to help analysts understand data clearly through graphs and plots. Visualizing data using WEKA is simple and can be done with the help of a box plot. It is done on the IRIS.arff dataset. In WEKA, you can represent data in so many ways. Some of them include pixel-oriented visualization, geometric representation, icon-based visualization, and hierarchical data visualization.

Sequential Pattern Mining Method

The sequential pattern mining method highlights patterns of ordered events that are in a database. A real-world example of a sequential pattern is a customer who buys a canon digital camera is likely to buy a color printer within a month. Sequential pattern mining is usually used for retail data to create promotions and place items on the shelf. Also, it can be used in other industries for target marketing.