Resident representative activity pattern recognition method and system based on LDA topic modeling

The invention discloses a resident representative activity pattern recognition method and system based on LDA topic modeling. The method and system are used for solving the problem that similarity of nominal sequences cannot be measured through a clustering algorithm. Through semantic conversion, re...

Full description

Saved in:
Bibliographic Details
Main Authors YE ZHENGHAO, LI DONGDING, SONG YUCHEN, HUANG DONGHAO, QIU SHURONG, LI DAWEI
Format Patent
LanguageChinese
English
Published 30.05.2023
Subjects
CALCULATING
COMPUTING
COUNTING
ELECTRIC DIGITAL DATA PROCESSING
PHYSICS
Online AccessGet full text

Cover

More Information
Summary:The invention discloses a resident representative activity pattern recognition method and system based on LDA topic modeling. The method and system are used for solving the problem that similarity of nominal sequences cannot be measured through a clustering algorithm. Through semantic conversion, resident questionnaire survey data is converted into a discretized activity sequence, the degree of confusion is calculated to determine the optimal theme number, and an LDA theme distribution model is adopted to obtain potential theme composition and distribution of a resident activity mode. And taking the obtained activity theme distribution as a new clustering unit, and applying a CURE algorithm to further extract a representative resident activity mode. According to the method, potential topics can be mined and extracted from a large amount of complex irregular trip chain information, and an activity mode with an obvious rule is obtained. The resident activity pattern recognition result can be used as an individu
Bibliography:Application Number: CN202310170758