HASTIE_NAIVEBAYES: from the Hastie_10_2.csv file obtained by the procedure described in https://scikit-learn.org/stable/modules/generated/sklearn.datasets.make_hastie_10_2.html, used both as test and training, obtains a success rate in the training of 89,29% and 87,13% in the test. These rates are slightly below that obtained by other procedures (see reference at the end), but they can be accepted, taking into account that a large number of the records do not meet the condition: y [i] = 1 if np.sum (X [i] ** 2)> 9.34 else -1 that is, there are erroneous data in the training.
The main difference in this model comparing with others is that in the statistical process, each field is sampled differently according to its contribution to the hit rate..
Resources: Java 8
The programs need to record files on disk:
Hastie_10_2Weighted88PerCentHits,txt, with all te records of training, adding 0 or 1 if the record have been considered hit or failures. This allows one manual inspectión and test the results.
HastieTestWithClassAsigned.txt tat is te output of the test files with the classes assigned acording wit the model.
For this reason, it is necessary to create a folder in the C: directory with the name of HASTIE_NAIVEBAYES and to which writing rights are assigned to the user who executes the program. Ultimately c: drives are write-protected. You can also change the directory and folder, but in that case you would have to change the references to C:\HASTIE_NAIVEBAYES in the .bat file.
Functioning:
Download Hastie_NaiveBayes.jar and Hastie_NaiveBayes.bat to any directory. The test file Hastie10_2.csv is downloaded to the C: directory.
Execute: Hastie_NaiveBayes.bat
The training and test file margins go as parameters, initially they are from register 0 to 9600 for training and from 9601 to 12000 for test, but they can be changed.
You can also use a test file that has the same structure as Hastie10_2.csv, in which the class goes in the first position of the record, by changing changing the assignment to he file and records margin in the only line in Hastie_NaiveBayes.bat.
On screen it offers the results and produces a file c:\HASTIE_NAIVEBAYES\HastieTestWithClassAsigned.txt with the classes assigned to the test file.
Cite this software as:
** Alfonso Blanco García ** HASTIE_NAIVEBAYES
References:
https://scikit-learn.org/stable/modules/generated/sklearn.datasets.make_hastie_10_2.html
Comparison:
Implementation of AdaBoost classifier