TY - GEN
T1 - Learning partial-value variable associations
AU - Ye, Nong
AU - Fok, Ting Yan
N1 - Funding Information:
The authors would like to thank the EPSRC and Dunlop Aerospace for financial support, Dr Toby Hutton and Dr Ron Fisher of Dunlop Aerospace for their guidance and assistance, and the technical staff of the Advanced Materials Research Group at the University of Nottingham.
Publisher Copyright:
© 2019 Association for Computing Machinery.
PY - 2019/5/10
Y1 - 2019/5/10
N2 - Existing machine learning and data mining techniques have difficulty in handling three characteristics of real-world data sets all together in a computationally efficient way: (1) different data types with both categorical data and numeric data, (2) different variable relations for different values of variables, and (3) unknown variable dependency. The Partial-Value Association Discovery (PVAD) algorithm handles these three characteristics of real-world data all together, and enables the discovery of partial-value and full-value variable associations as well as univariate and multivariate effects of variable values. This paper illustrates the PVAD algorithm using a small data set, and presents the new development of a computationally fast method to discover the longest associations with one conditional variable value or one associative variable value.
AB - Existing machine learning and data mining techniques have difficulty in handling three characteristics of real-world data sets all together in a computationally efficient way: (1) different data types with both categorical data and numeric data, (2) different variable relations for different values of variables, and (3) unknown variable dependency. The Partial-Value Association Discovery (PVAD) algorithm handles these three characteristics of real-world data all together, and enables the discovery of partial-value and full-value variable associations as well as univariate and multivariate effects of variable values. This paper illustrates the PVAD algorithm using a small data set, and presents the new development of a computationally fast method to discover the longest associations with one conditional variable value or one associative variable value.
KW - Data mining
KW - Machine learning
KW - Partial-value data association
UR - http://www.scopus.com/inward/record.url?scp=85069781326&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85069781326&partnerID=8YFLogxK
U2 - 10.1145/3335484.3335497
DO - 10.1145/3335484.3335497
M3 - Conference contribution
AN - SCOPUS:85069781326
T3 - ACM International Conference Proceeding Series
SP - 24
EP - 28
BT - ICBDC 2019 - Proceedings of 2019 4th International Conference on Big Data and Computing
PB - Association for Computing Machinery
T2 - 4th International Conference on Big Data and Computing, ICBDC 2019
Y2 - 10 May 2019 through 12 May 2019
ER -