C4.5

所属分类:Java编程
开发工具:Java
文件大小:1KB
下载次数:124
上传日期:2008-04-21 12:06:48
上 传 者xinyuanwo
说明:  C4.5算法有如下优点:产生的分类规则易于理解,准确率较高。其缺点是:在构造树的过程中,需要对数据集进行多次的顺序扫描和排序,因而导致算法的低效。此外,C4.5只适合于能够驻留于内存的数据集,当训练集大得无法在内存容纳时程序无法运行。
(C4.5 algorithm has the following advantages: the classification rules easier to understand, accurate and a higher rate. Its shortcomings are as follows: in the tree structure, the need for a number of data sets the order of scanning and sorting, thus leading to inefficient algorithms. In addition, C4.5 can only be applied to the presence of a data set in memory, when the training set too great to accommodate in memory when the program can not run.)

文件列表:
C4[1].5 java 源代码\c45\C45Driver.java (1414, 2002-05-21)
C4[1].5 java 源代码\c45 (0, 2002-05-22)
C4[1].5 java 源代码 (0, 2008-04-21)

To compile: Set $MLJDIR = location of mlj dir. (Ex. /home/mad1333/mlj) Next, go to the shared directory and run make c45. To run ID3 Driver: Then from the shared directory run: java -cp .:../id3:../c45 C45Driver ../../data/$DATASET where $DATASET = the name of the data set you want to run. This will run the Driver over $DATASET.names, $DATASET.test, and $DATASET.data in the ../../data directory. A second argument can be supplied to change the loglevel and global loglevel to a different setting. Example: java -cp .:../id3:../c45 C45Driver ../../data/$DATASET 4 This will change the loglevel to 4 and produce more output. Not supplying a value results in a default value of 0. A third option allows switching between tree display specifications as they are in the MLC binary and specifications as they are in the MLC code. To switch between these settings, add true or false as a third setting. Example: java -cp .:../id3:../c45 C45Driver ../../data/$DATASET 4 false False sets to the source specifications and true sets to the binary specifications. The default is true.

近期下载者

相关文件


收藏者