2.1 Overview
2.2 Total Operation in 2 Clicks
2.3 Predict What?
2.4 Example: Intel stock Value
2.4.1 Phase 1: Intel Raw Data
2.4.2 Operating the PredictorPro

2. Total Operation in 2 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.

The operation of the PredictorPro is:

Link; and
Run.

Other than preparing the input file, the PredictorPro is completely �link and run�. No typing is required to operate this Predictor.

2.1 Overview

Assume you have a large database, say you have collected 30 years of data, and you want to make intelligent decisions based on the huge amounts of information in your database.

Here is your difficulty: you are overwhelmed by your terabyte data and lose a sense of the direction in the data.

What are you going to do? Order Attrasoft PredictorPro.

The prediction process has two phases:

Phase 1: from your historical database, you will determine a data set, which consists of the relevant variables for the prediction. Then get the data for these variables, if necessary, preprocess them such as calculating the moving-average.
Phase 2: Prepare all your data in one file; then in 2 clicks, you will have your answer instantaneously:

Prepare Input file: this is all the data you have collected about your problem and put into one file; this file will teach the PredictorPro your problem and the value of your experience.

The first click ----- link the data to the PredictorPro;

The second click ----- get the prediction(s) to your question(s);

Now the PredictorPro becomes one of your staff, and is ready to serve you. The predictions will be in a file opened automatically after the second click.

2.2 Total Operation in 2 Clicks

Assuming in phase 1, you have accumulated the data.

Formulate your raw data into a file, the input file;
Link the input file to the PredictorPro by clicking "File/Open" and open the input file (first of 2 clicks).
Once the file is in the PredictorPro, click a command, say "Real/+ Exponential" (second of 2 clicks).

PredictorPro presents you with all possible predictions and how valuable the PredictorPro thinks each prediction is via a confidence number or rating.

The confidence number is similar to the number in an Internet Search Engine: the higher that number, the more confidence the PredictorPro has in that prediction.

2.3 Predict What?

The PredictorPro predicts the values of the next period or next several periods. Take stock data as an example,

If you have daily data, you will predict next day's data;
If you have weekly data, you will predict next week's data;
If you have monthly data, you will predict next month's data;
...

2.4 Example: Intel stock Value

In this section, we want to predict the Intel stock value for 10/98.

2.4.1 Phase 1: Intel Raw Data

1. Data Set

A data set is your choice of the variables, which you think will give you the best prediction. We will choose the Intel stock as the sole factor for the prediction, i.e., the data set is the Intel stock. (This is, of course, only for the purpose of demonstrating how to operate the PredictorPro. Attrasoft does not recommend you use any particular data set for a stock.)

2. Data Collection

Free from the Internet, we can collect the following data. In chapter 5, we will show you how to do this.

Intel Corporation (INTC)
Monthly prices (Nov 1991 to Nov 1998)
DATE       HIGH     LOW     CLOSE   VOLUME        5m Avg.
Nov-91     5.531     4.969     5.125       275020000
Dec-91      6.156     5.031     6.125       260759200
Jan-92       7.875     5.906     7.484       589994400
Feb-92      8.594     7.375     7.938       340300800
Mar-92      8.438     6.438     6.906       424595200     6.7156
Apr-92      7.406     6.125     6.688       443011200     7.0282
May-92     6.844     5.813     6.25         300004000     7.0532
Jun-92       7.219     5.969     7.125       346980000     6.9814
Jul-92        7.531     6.844     7.469       291545600     6.8876
�

May-98    85.438     71.25     71.438     271927700     80.2004
Jun-98      77.625     65.656   74.125     360081300     78.8254
Jul-98       88.25       72.25     84.438     396391200     77.7754
Aug-98     92.625     70.938   71.188     440267900    76.4004
Sep-98     89            69.75     85.75       369103300     77.3878
===================================================
Oct-98     90.813    75.813   89.188     393720800     80.9378

We will split the data into two parts:

Nov-91 through Sep-98 data will be used to train the PredictorPro,
Oct-98 data will be used to test the PredictorPro.

Note: chapter 5 will show you how to collect free stock data from the Internet.

3. Preprocessing

To remove the noise, we will use the 5-month moving average, which is listed in column 6. Because this is monthly historical data, we will predict the 5-month moving average of Intel for the next month.

2.4.2 Operating the PredictorPro

Step 1. Prepared Input file

From the above raw data, set up the Intel file (5-month moving-average) in the following manner:

The input file is:
*
Purpose: To predict next month Intel.
Data set: 5-month moving average of Intel, Nov-91 through Sep-98
Next 5-month moving average:
10/98 80.9378
*
1
6.7156
7.0282
7.0532
6.9814
6.8876
�
80.2004
78.8254
77.7754
76.4004
77.3878

This Intel input file has three sections:

Remark section:

PredictorPro

Number of variables/attributes:

Data :

Figure 2. The Intel input file.

Step 2. Link the data to the PredictorPro

The above input file, "D_int2a.txt", is in the default directory of the software. Click "File/Open" and open this file. This is the first of 2 clicks. The linking is completed ... just that easy.

To check if the file is linked correctly, click �Data/Link� (see Figure 3)., and a dialog box will appear (see Figure 4).

Figure 3. Click command: "Data/Link".

Figure 4. Link the input file to the PredictorPro

If the PredictorPro is linked to the data files, then the first textbox in Figure 4 should be �D_intc2a.txt�, which is the input file. We will explain the other three textboxes later.

Step 3. Click command: "Real/-- Exponential"

Figure 5. The "Real" Menu.

Figure 5 shows the "Real" menu. "Real" means real number prediction. The other prediction is "Integer" prediction. Click "Real/-- Exponential" (second of 2 clicks) to get the answer. The running takes no time, and the following prediction file is opened automatically (See Figure 6):

=================== Beginning =====================
Possibility              Confidence*Probability
81.2484       6400
77.8113       10344
76.0928       4235
84.6855       273
88.1226       1
86.404       33
79.5299       168
74.3742       8320
82.9669       261
---------------------------------------------------
Weighted Average
77.476
Highest Probability
77.8113 10344
Error of each number
0.859271
=================== End ==========================

As we have shown in the earlier this section, the correct answer is:

DATE 5 month moving average
Oct-98 80.9378

Ignoring results below 1000, the prediction is

Value                                Probability Confidence
81.2484 ± 0.859271      0.218           6400
77.8113 ± 0.859271      0.353           10344
76.0928 ± 0.859271      0.144           4235
74.3742 ± 0.859271      0.283           8320

where 0.859271 is given by the last line of the output file. The prediction is:

Weighted Average = 77.476 ± 0.859271
Highest Probability = 77.8113 ± 0.859271 (35.3 % probability)

This prediction is very close to 80.9378. The probabilities of the predictions are calculated as follows:

6400 / (6400 + 10344 + 4235 + 8320) = 0.218
10344 / (6400 + 10344 + 4235 + 8320) = 0.353
...

Figure 6 The results of clicking "Real/-- Exponential" (second of 2 clicks)

Online Order/Online Fax order:
Attrasoft PredictorPro for Windows
Version 2.8 ($494.99 + $5 US Shipping and Handling)

Mail Order: 499.99 (S&H included)

PredictorPro
Attrasoft, Attn.: Gina
P. O. Box 13051
Savannah, GA. 31406, USA

Send questions or comments to:

webmaster@attrasoft.com

or contact us at:
Attrasoft, P. O. Box 13051, Savannah, GA. 31406, USA