|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2009年04 期
| 標(biāo)題 |
| 改進(jìn)的屬性約簡算法及其在決策樹中的應(yīng)用(18 卷) |
| 英文標(biāo)題 |
| Improved attribute reduction algorithm and its application to decision tree |
| 摘要 |
| 針對傳統(tǒng)的基于區(qū)分矩陣屬性約簡算法在構(gòu)建區(qū)分矩陣時生成很多冗余元素，空間和時間性能上不夠理想的缺點(diǎn)，利用動態(tài)數(shù)組及新的約簡過程進(jìn)一步改進(jìn)，最后把新的屬性約簡算法應(yīng)用到?jīng)Q策樹的構(gòu)建上，實(shí)驗證明，該算法能夠提高決策樹分類性能。 |
| 作者 |
| 新聞作者:劉香美，樓新遠(yuǎn) |
| 關(guān)鍵字 |
