DEVELOPMENT OF APPLICATION OF BAYESIAN NETRWORKS AT DATA MINING
Data mining is a process of extracting data to get knowledge from large data repositories. One of method of data mining is classification that is process of seeking model of classification that able to differentiate its class label object. Bayesian Networks is one of technique that able to use to bu...
Saved in:
| Main Author: | |
|---|---|
| Format: | Theses |
| Language: | Indonesia |
| Online Access: | https://digilib.itb.ac.id/gdl/view/8093 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Institution: | Institut Teknologi Bandung |
| Language: | Indonesia |
| Summary: | Data mining is a process of extracting data to get knowledge from large data repositories. One of method of data mining is classification that is process of seeking model of classification that able to differentiate its class label object. Bayesian Networks is one of technique that able to use to build classification model. Component Bayesian Network compose two component that is DAG structure depicting causality relationship among data attribute and a containing tables of conditional probability based on previous attribute called CPT. TPDA (Three Phase Dependency Analysis) is one of algorithm to build structure of Bayesian Networks. This algorithm builds structure of Bayesian Networks based on value of information is found on two node. Ever greater of value of information can flow of between two node hence is ever greater of opportunity to connect of the nodes. TPDA consist of three phase. First phase is drafting that is phase to determine early graph to connect a pairs of node. Second phase is thickening that is phase that add edge to current structure when a pairs of node can not separated. Third phase is thinning that is phase where each of edge is examined and it will be removed if two node if found to be separated. Based on result of performance evaluation of class inference using structure of Bayesian Network yielded by application in this thesis, seen that level of inference accuration is high enough. But when structure of Bayesian Networks yielded by application in this thesis is compared with structure yielded by BN PowerConstructor there are difference between of them. On that account, the application has developed in this thesis still need repaired to get same structure, because BN PowerConstructor which is developed by implementing TPDA algorithm have proven its truth. <br />
|
|---|