Data Science 6

Cluster Analysis

A method of finding groups of objects that are similar within the same group and different from those in other groups.

Intra-cluster distances

Distances between objects within the same cluster, which are minimized in cluster analysis.

Inter-cluster distances

Distances between objects in different clusters, which are maximized in cluster analysis.

Segmentation

The process of dividing customers into groups based on similarities in behaviors, attributes, or patterns.

Prototype customers

Descriptions of typical customers in a segment, helping to tailor marketing messages.

Clickstream profiling

The analysis of patterns from user clicks on a website to understand different customer behavior types.

Document clustering

Grouping documents into clusters based on similar topics, facilitating topic identification and summarization.

Partitional Clustering

Clustering into non-overlapping subsets where each instance is part of exactly one cluster.

Hierarchical Clustering

Clustering that organizes data into a set of nested clusters structured as a tree.

Density Based Clustering

Clustering where examples in dense areas form a cluster, while those in sparse areas do not.

k-Means Clustering

A method that partitions a dataset into k clusters based on proximity measures such as Euclidean distance.

Euclidean distance

A measure of distance between two data objects, commonly used in clustering to determine similarity.

Association Analysis

A technique that measures the strength of co-occurrence between items to find patterns in data.

Market Basket Analysis

A method in Association Analysis to find co-occurrences of items in retail transactions.

Cross-selling

Recommending additional products to customers based on their previous purchases.

Confidence of the Rule

The likelihood that the consequence of an association rule occurs given that the antecedent occurs.

Antecedent of the Rule

The item or items that, if present in a transaction, suggest a strong likelihood of another item being present.

Consequent of the Rule

The item or items that are predicted to occur given the presence of the antecedent.

Frequent Item Sets

Combinations of items that appear frequently in transaction data, essential for generating association rules.

Transaction Format

The specific data organization required by association algorithms for effective analysis.

Interest Measure

A criterion used to filter generated association rules to ensure they are meaningful.