liveBook · Manning

Support threshold

As explained in Data Without Labels

A support threshold is the minimum frequency or proportion required for a sequence pattern to be considered frequent.

A support threshold sets the minimum frequency required for a sequence pattern to remain frequent. In sequential pattern discovery, it filters candidate sequences before longer combinations are evaluated. section 4.7.1

Counting ordered purchases

Using Milk and Eggs, a support threshold of 2 retains patterns occurring at least twice. Milk -> Eggs qualifies for sequence IDs 1001, 1002, 1003, and 1004; Eggs -> Milk qualifies for 1001 and 1002. Ordering is checked by comparing element IDs within the same sequence ID. section 4.7.1

A Python workflow can pass the threshold directly to SPADE through pycspade; support=0.6 runs the mining process and prints its results.

How Data Without Labels covers it

How deep it goes
Works through sequence IDs, element IDs, ordered pairs, and joins, then stops short of threshold-result interpretation.
Examples
Uses Milk-and-Eggs purchase sequences, support threshold 2, and a pycspade call with support=0.6.
What it assumes
Assumes sequence IDs and element IDs; establishes vertical sequence representation before counting patterns.
What you can do after
Apply it via pycspade by setting the support parameter before running SPADE.
How it connects
Connects sequential pattern discovery with vertical data format, sequence mining, and SPADE.
Cover of Data Without Labels
liveBookmanning.com
Read Data Without Labels on liveBook Get the book at manning.com

This explainer was generated by Manning's knowledge graph from the text of Data Without Labels. The section links point to the parts of the book it draws on.