Mining High Average Utility Temporal Patterns on Time Interval-Based Data
碩士 === 國立交通大學 === 資訊科學與工程研究所 === 108 === In recent year, mining high utility temporal pattern becomes an important topic in the fields of data mining. Previous researches of utility temporal pattern mining focused on the relations between the events. However, the number of event occurrence of the hi...
| Main Authors: | , |
|---|---|
| Other Authors: | |
| Format: | Others |
| Language: | en_US |
| Published: |
2019
|
| Online Access: | http://ndltd.ncl.edu.tw/handle/79gzng |
| Summary: | 碩士 === 國立交通大學 === 資訊科學與工程研究所 === 108 === In recent year, mining high utility temporal pattern becomes an important topic in
the fields of data mining. Previous researches of utility temporal pattern mining focused
on the relations between the events. However, the number of event occurrence of the high
temporal pattern may very high. In the real world application, people tend to retrieve
low event occurrence and high utility temporal patterns to get high profit. To solve the
problem, we raise the methodology of mining high average utility temporal pattern. In
this paper, we proposed two pruning strategies to reduce the large search space. HAUTPMiner algorithm is proposed to efficiently mine the high average utility temporal pattern.
A series of experimental results show that the two pruning strategies could efficiently and
effectively retrieve the high average utility sequences.
|
|---|