Attrasoftfor Windows 95/98 Version 4.0 (7/1999) |
||||||||||
|
All Attrasoft products are Y2K Compliant! 2. Total Operation in
2-4 Clicks
The detailed user's guide will be given in chapter 3 and 4. In this chapter, we will show you the minimum information to get started. We will use two examples, Cancer diagnosis and character recognition. 2.1
Total Operation in 2-4 Clicks
Prepare all your training data in one file and all your recognition data in another file, then in 2-4 clicks, you will have your answer instantaneously. The procedure is Step 1. Put the training data into the training file; Put your recognition data into the recognition file; together you prepare two files. Step 2. Link the two files to the PolyNet by clicking "Data/Link". Step 3. Once the files are in the PolyNet, set up symmetry and output format for PolyNet if necessary; Step 4. Click 'D' button (Decimal/Distribution command) for a distribution or 'C' button for a classification (Decimal/Classification command). 2.2 Example: Wisconsin Breast Cancer Database The question is do you have cancer or not? You must have
1. Sources: This breast cancer database was obtained from Dr. William H. Wolberg, University of Wisconsin Hospital, Madison. Please see: O. L. Mangasarian and W. H. Wolberg: "Cancer diagnosis via linear programming", SIAM News, Volume 23, Number 5, September 1990, pp. 1 & 18. 2. Database Information
Malignant: 241 (34.5%) #
Attribute
Domain
1.
Sample code number
id number
2.
Clump Thickness
0 - 9
11. Class: (1 for benign, 3 for malignant) Note that the original data is the range 1 - 10. The data has been changed
to 0 -9.
4. Database Below are the first 10 instances: 4 0 0 0 1 0 2 0 0 1
Below are the last 20 instances, which will be used to test the PolyNet's accuracy: Question Answer 1 0 0 0 1 0 0 0 0 1
2.2.2 Training and Recognition Files Cancer is usually evaluated and diagnosed by the above test variables, which is called a data set. Assume we have the cancer test data which doctors/laboratory technicians have already accumulated into a Cancer Database. The question is: from this data, do the following two patients have cancer (1 for benign, 3 for malignant)? 10,7,7,6,4,10,4,1,2,? 6,1,1,1,2,1,3,1,1,? The 683 rows of data in the Wisconsin Breast Cancer database will be divided into two parts:
All data files must be in the text format and each has three sections: remark, number of columns, and data. Step 1. Training file and Recognition file The first file, ... in this case ... Cancer-database file, is used to
train the PolyNet. Let the name of the file be "cancer1a.txt" and
it will look like this:
10 4 0 0 0 1 0 2 0 0 1
============= End =============
Figure 2. The Cancer-Database Training File.
This Cancer-Database file has three sections:
============== Beginning =============
1 0 0 0 1 0 0 0 0 x
============= End ============= This recognition file has 2 sections:
Figure 3. The Recognition file. Step 2. Link the data to the PolyNet (1 or 2 click(s) ) To link the cancer data, all you do is to click "Example/Cancer Database". The data will be linked and the training file, cancer1a.txt, and the recognition file, cancer1b.txt, will be opened at this point. In general, save the training and recognition files in a directory.
Click "File/Open" to open one of the two files. This will lead the
PolyNet
to the directory. Click "Data/Link" (see Figure 4) and a dialog
box will ask you for the following (see Figure 5):
1. Training File Name;
Figure 4. Click command: "Data/Link".
Figure 5. Link the data files to the PolyNet
To link the data files, in the first 2 textboxes (Figure 5), enter
Step 3. Once the files are in the PolyNet, set up symmetry and output
format for PolyNet if necessary;
Do nothing here.
Figure 6. "Decimal" Menu. Figure 6 shows the "Decimal" menu. Click "Decimal/Classification" to get the answer. The following output file is opened instantaneous: ============== Beginning ============= 100010000x
999949996x
499939452x
... ============= End ============= In the first output: 1000100001 3523 1 2.3 The '5 by 7' Character Recognition Problem In this section, we will introduce the '5 by 7' character recognition
problem.
2.3.1 Step 1: Training and Recognition Files Characters 0, 1, 2, 3, ... look like this:
01100 01110
01000 00000 0 00100
00100is in class "1": 01000 00000 0. The training file is a list of images like: 01000 00000 0 00100
00100
xxxxx xxxxx x The neural net has to determine that for the above pattern, xxxxx xxxxx x is 01000 00000 0. The recognition file is a list of images like: xxxxx xxxxx x 00100
Example/5x7 Character. This action will generate 46 training-characters, and 10 recognition-characters. The two files will also be opened by this click. 2.3.2 Step 2: Link the Two Files You can link the two files by using 'Data/Link' and enter the file names. The default names are Training file:
chara.txt
When you click "Example/5x7 characters", the training and recognition files are generated and linked to the PolyNet. Before you can train the network, you have to set up the network. The setup deals with:
To set up the symmetry, use the command "Data/Symmetry". The character recognition problem has 2-dimensional x- and y-translation symmetries. Symmetry very is important for training the network. When you click "Example/5x7 characters", the output format and the symmetry
are setup for you by PolyNet. Click "Data/Symmetry" to see the symmetry
settings. We will explain this later.
Now that the network is initialized and the data files are linked, it is the time to run. There are two commands for classification:
Click: Decimal/Classification. The results from this command are obtained
immediately. The output file, which contains the results, will be opened
at this point. It looks like this:
xxxxxxxxxxx
10000000000. The classification is 100% correct in this example.
Online Order
or if you prefer, Online Fax order:
Mail Order: $129.99 (S&H included) PolyNet 4.0
Send questions or comments to: webmaster@attrasoft.comor contact us at: Attrasoft, P. O. Box 13051, Savannah, GA. 31406, USA Copyright © 1998 Attrasoft, Inc. All rights reserved. |