Text mining with minimal human supervision
In a National Basketball Association match, head coaches often have to make good and timely decisions in the best interest of his team. To make good decisions, it is important that coaches know and recognize his players’ past performances and records.This report explores the various aspects in the c...
Saved in:
| 主要作者: | |
|---|---|
| 其他作者: | |
| 格式: | Final Year Project |
| 語言: | English |
| 出版: |
2011
|
| 主題: | |
| 在線閱讀: | http://hdl.handle.net/10356/44614 |
| 標簽: |
添加標簽
沒有標簽, 成為第一個標記此記錄!
|
| 機構: | Nanyang Technological University |
| 語言: | English |
| 總結: | In a National Basketball Association match, head coaches often have to make good and timely decisions in the best interest of his team. To make good decisions, it is important that coaches know and recognize his players’ past performances and records.This report explores the various aspects in the creation and implementation of the system. The main objective of this project is to develop a system, NBA Automated Extraction System (NAXS), which uses the text mining technology and it follows the manual approach of identifying data patterns. The system automatically crawls the NBA Web site to search for games as specified by the user and extract useful information from these games.
The proposed system was evaluated for the precision of the extraction procedure though various tests. These tests comprise of data taken from 3 months and include the extraction of i) The number of games as well as ii) The extraction of individual player’s statistics. All the tests performed well with each having a percentage of above 99%. The average accuracy of the extraction procedure of NAXS based on the data taken from 3 months is 99.80%.
In conclusion, NAXS proved to be almost as efficient as counting the data manually and the automation process is also much faster as compared to the manually counting process. |
|---|