Attrasoft Predictor/User's Guide/Chap. 4

N-trend means the Predictor looks at N rows of data at a time. This is one of the three specifying parameters used by the Predictor. The N-trend is a user-selected variable. If you do not make a choice, the default value is 5.

To change the N-trend (See Figure 4):

click "Data/Link";
Go to "Enter trend", and enter your number.

Under similar prediction confidence numbers, the longer the Trend, the more accurate the prediction will be.

You have to balance the two sides of this parameter. On one hand, a 10-trend certainly provides a more accurate prediction than a 5-trend. On the other hand, a 10-trend prediction requires a much larger volume of data. One way to increase N is to use preprocessed data. Using the moving averages can increase N significantly.

Precision Level

The Precision-level determines the error of each variable. The higher the Precision is, the lower the errors are. This is one of the three specifying parameters used by the Predictor. The Precision level is a user-selected variable. If you do not make a choice, the default value is 10.

This parameter will directly change the errors in the output file. To change the Precision level (See Figure 4):

click "Data/Link";
Go to "Enter precision level", and enter your number.

You have to balance the two sides of this parameter. On one hand, a 10-level prediction certainly provides more precision than a 5-level prediction. On the other hand, a 10-level prediction requires a much larger volume of data. One way to increase the precision level is to use preprocessed data. Using the moving average can increase the precision level significantly.

Trend, Precision and Confidence

In general, the longer the trend, the more accurate the prediction will be. The higher the precision, the lower the error(s) will be.

Between these two factors, the Trend is more important than Precision. If possible, always goes for the maximum Trend.

Here is how to set the Trend: as you increase the Trend from low to high, initially, you can see the confidence (rating) increase. If you keep increasing the Trend, for most problems, at certain point the confidence begins to drop. You should establish an acceptable confidence level and that will tell you when to stop.

Example "Example/Intel 2"

Click "Example/Intel 2" to generate the example, which predicts the 5-month moving average of Intel on October, 1998.

Set Trend = 10 and click "Real/+ Exponential" and the output is:

Possibility Confidence*Probability

Can not make a Prediction.

To get a prediction, you can:

(1) Add more data;

If you do not have more data; then click 'Data/Link' and:

(2) Reduce Trend; or/and

(3) Reduce Precision level;

You may also consider:

(4) Reduce the number of variables in your data set.

See the User's Guide

---------------------------------------------------

Ops, Trend = 10 is too high. Set the Trend to 9, and click "Real/+ Exponential" again to get:

Possibility Confidence*Probability

77.8113 16384

---------------------------------------------------

Weighted Average

77.8113

Highest Probability

77.8113 16384

Error of each number

0.859271

Therefore, the maximum Trend is 9.
In general, the higher the Trend is, the lower the Precision level.

Roles of the Four Parameters

The four parameters have the following impact on the prediction:

Data volume: one year's data will produce more accurate results than one month�s data;
Number of variables in the data set: more variables requires more data;
Trend: a longer trend requires more data than a shorter trend;
Precision level: a higher precision prediction requires more data than a lower one;

If you run the Predictor, and get too much information, you can do one or more of the following:

Reduce the data volume;
Increase the number of variables in the data set;
Increase the Trend;
Increase the Precision level.

If you run the Predictor, and get no results, you can do one or more of the following:

Data preprocessing;
Increase the data volume;
Reduce the number of variables in the data set;
Reduce the Trend;
Reduce the Precision level.

Data preprocessing can improve the result significantly. There are noise fluctuations in the data. These noises can be removed from the data by a simple procedure like introducing the moving averages. However, if the noise fluctuation is left in the data, the Predictor will think these noise fluctuations are meant to be learned and will treat the noise fluctuation as well-defined patterns in the data. As a result, it will shorten the Trend and reduce the Precision. As we will see, preprocessing (such as the moving average) can be achieved in two steps in Excel, which takes only a few seconds to accomplish.

Data selection

Data preparation for a prediction is the most important factor in the prediction or forecast. Data preparation means two things:

data set; and
the number of rows.

Your data preparation directly influences the results. In this section, we will show you how to prepare data.

The better you choose variables, the better the potential for the prediction. This is the single most productive area to work on. You can increase the information content of your data in several ways:

Adding variables: the first approach is to add variables to the data set. Assuming that a new variable captures some aspect of the problem not already included in the other variables you've chosen to be in the data set, the information content is increased. On one hand, obviously, you should continue to do so until you think all aspects of the problem are covered. Otherwise, the data set will not be complete. On the other hand, each additional variable increases the complexity of the problem exponentially. To compensate for this, more data is required. If your data is limited, then your ability to make a prediction is limited.

Data Preprocessing: This approach transforms the existing data into another form. Typically, this is just a mathematical manipulation of the data. A moving average, for example, is a mathematical manipulation of the underlying price variable. This approach may allow us to extract information more effectively or more efficiently. For stock prediction, we always recommend you to use the moving averages of the underlying stock price.

Removing Unnecessary Variables: This approach removes data that does not increase the information content. Overloading the Predictor with useless or redundant information will increase the complexity of the computation exponentially and will require more data. If you do not have enough "rows" for the additional variables, the quality of the prediction can be reduced. Averaging Intel and Microsoft stock prices, for example, is a mathematical manipulation, which covers both factors with only one number.

Increase Data Volume: This approach increases the data volume. Two years of data, for example, is better than one year of data, if the conditions in these two years are basically the same. In reality, there is a limit to the data volume. For the Attrasoft Predictor, always go for the maximum amount of available data. The cost of increasing data volume in computation time is almost 0. The data volume is not a factor in the Attrasoft Predictor, for the software is designed for terabyte processing.

Using similar data: It is quite possible that sometimes you have no data. In this case, you might consider using similar data. We will present you with an example in which, despite not having Intel data, we still want to predict Intel stock by substituting Intel with Microsoft, because Microsoft stock is similar to Intel stock. As you will see, the Predictor will still be able to make a direct hit.

Once again, data set selection is your job. This is the single most productive area to work in, regardless of the technology employed for your data-mining effort. Your largest contribution to your prediction is right here.

Using Excel as Your Text Editor

Almost any word processor can be used as the text file editor, Microsoft Word, WordPerfect, Windows Notepad, ..., as long as you save the files in text format. We recommend you to use Microsoft Excel as a text editor. First of all,

Excel opens data in text files directly; and
Excel can save files into text format directly.

Secondly, Excel gives you the power to manipulate the data easily. The functions in Excel you might use repeatedly are:

Opens text data in any format (numbers separated by a space, tab, comma, ... )
Moving average;

Sorting;

Selecting a column or several columns from a data file;
. . .

Data Preprocessing Using Excel

Data preprocessing transforms the existing data into another form. Typically, this is just a mathematical manipulation of the data. A moving average, for example, is a mathematical manipulation of the underlying price variable. This approach may allow us to extract information more effectively or more efficiently.

Moving-Average

When dealing with data with noise fluctuations, like stock market prediction, always use the moving averages of the underlying variables.

If there are noise fluctuations in the data, these noises can be removed from the data by a simple procedure like moving average. However, if the noise fluctuation is left in the data, the Predictor will think these noise fluctuations are meant to be learned and will treat the noise fluctuations as well-defined patterns in the data. As a result, it will shorten the Trend and reduce the Precision of a prediction.

It is an easy job to convert the original data to the moving average data. There are only two steps to convert the data:

Compute the first moving average;
Use "Edit/Fill/Down" to fill in the rest of the moving averages.

Now we will show these two steps in detail. Assume we want to calculate the 5-month moving average of the SP 500. Figure 10 shows the data.

Figure 10. SP 500 data in Excel.

The data is as follows:

A B C

3
DATE CLOSE

4

Nov-91

375.22

5

Dec-91

417.09

6

Jan-92

408.78

7

Feb-92

412.7

8

Mar-92

403.69

9

Apr-92

414.95

10

May-92

415.35

. . .
Step 1: at the C8-cell, enter "=SUM(B4:B8)/5".

Here "SUM(B4:B8)" means B4 + B5 + B6 + B7 + B8 in Excel language. Please see Figure 11.

Figure 11. First moving average of the SP 500 in C8-cell.

Hit enter and the first moving average is calculated. In this case, the first moving-average is 403.496 (See Figure 12).

Figure 12. First moving average of the SP 500 is in C8-cell.

Step 2: Highlight the cells: C8, C9, C10, ... , (See Figure 12) then click: "Edit/Fill/Down".

Now the moving average computation is completed. The result looks like Figure 13.

Figure 13. The rest of the moving averages of the SP 500 in C-column.

Example of Preprocessing

In this section, we will compare the result of using the original data and the result of using the moving average data (preprocessed data).

Two examples will be used. Both are packed in the Predictor, which can be generated by clicking "Example/Intel 1" and "Example/Intel 2".

Using Original Data

1. Click "Example/Intel 1" and the original data will appear:

5.125

6.125

7.484

7.938

6.906

...

71.438

74.125

84.438

71.188

85.75

The target (for Intel on Oct-98) is 89.188.

2. Click "Real/-- Exponential" and the following results will appear:

Possibility Confidence*Probability

73.138 128

95.1938 256

---------------------------------------------------

Weighted Average

87.8418

Highest Probability

95.1938 256

Error of each number

2.20558

Using the weighted-average, the answer is: 87.8418 ± 2.20558 = 87.8 ± 2.2,
The confidence number, 256, is extremely low, indicating the Predictor does not have much confidence in the data at all. Assume we use 500 as the minimum acceptable confidence-level, then the above prediction would not be used.

To get a prediction, we have to:

Reduce the Trend; or
Reduce the Precision level.

Using the 5-Month Moving Average

1. Click "Example/Intel 2" and the 5-month moving average data will appear:

6.7156

7.0282

7.0532

6.9814

6.8876

...

80.2004

78.8254

77.7754

76.4004

77.3878

The target (for the 5-month moving average of Intel on Oct-98) is 80.9378.

2. Click "Real/-- Exponential" and the following results will appear:

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

The confidence number is 10344 in this example, instead of 256 in the last example, which indicates the Predictor is far more confident in this case. The prediction, 77.476 ± 0.859271 is close to 80.9378.

Example of Poor Data

Again we want to predict the 5-month moving average of the Intel for Oct-98, which is 80.9378.

Assume the only data we have for Intel is:

DATE HIGH LOW CLOSE VOLUME 5m Avg

Jan-98 82.875 70.5 81 339559700

Feb-98 95.625 82.188 89.688 333269000

Mar-98 91.375 74 78.063 425380400

Apr-98 85.063 71.313 80.813 386529700

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

This will not be enough for the prediction. To get around this problem, we find:

(1) Microsoft and Intel are similar; and
(2) We do have Microsoft data, as listed below:

Monthly prices (Nov 1991 to Nov 1998)

DATE HIGH LOW CLOSE VOLUME 5m Avg

Nov-91 8.479 7.563 8.104 221809200

Dec-91 9.333 8 9.271 202192800

Jan-92 11.104 9.125 10.021 367232400

Feb-92 10.75 9.417 10.292 348507600

Mar-92 10.854 9.729 9.875 301068000 9.5126

Apr-92 10.771 8.917 9.188 534740400 9.7294

May-92 10.188 9.104 10.083 267308400 9.8918

Jun-92 10.375 8.219 8.75 394200000 9.6376

Jul-92 9.344 8.188 9.094 311299200 9.398

Aug-92 9.375 8.5 9.313 222331200 9.2856

Sep-92 10.25 9.156 10.063 250730400 9.4606

Oct-92 11.281 9.469 11.094 286680000 9.6628

Nov-92 11.875 10.906 11.641 306790400 10.241

Dec-92 11.75 10.656 10.672 283332000 10.5566

Jan-93 11.75 10.563 10.813 352832800 10.8566

Feb-93 11.281 9.594 10.422 387085600 10.9284

Mar-93 11.781 10.156 11.563 299489600 11.0222

Apr-93 11.844 9.969 10.688 324658400 10.8316

May-93 11.938 10.563 11.578 326112000 11.0128

Jun-93 12.25 10.844 11 304052800 11.0502

Jul-93 11.063 9 9.25 488353600 10.8158

Aug-93 9.906 8.797 9.391 467870400 10.3814

Sep-93 10.531 9.156 10.313 290615200 10.3064

Oct-93 10.719 9.75 10.016 342688000 9.994

Nov-93 10.375 9.5 10 241628800 9.794

Dec-93 10.813 9.906 10.078 233696000 9.9596

Jan-94 10.875 9.906 10.641 308412800 10.2096

Feb-94 10.719 9.75 10.313 342180000 10.2096

Mar-94 11.156 9.969 10.594 344377600 10.3252

Apr-94 11.906 10.25 11.563 467880000 10.6378

May-94 13.441 11.359 13.438 407473200 11.3098

Jun-94 13.656 12.313 12.906 609798800 11.7628

Jul-94 12.938 11.719 12.875 381991600 12.2752

Aug-94 14.813 12.875 14.531 322749600 13.0626

Sep-94 14.563 13.719 14.031 241212400 13.5562

Oct-94 15.938 13.469 15.75 301788800 14.0186

Nov-94 16.281 15.188 15.719 256071600 14.5812

Dec-94 16.063 14.688 15.281 264838400 15.0624

Jan-95 16.313 14.563 14.844 277902400 15.125

Feb-95 15.813 14.594 15.75 249664400 15.4688

Mar-95 18.531 15.688 17.781 353931200 15.875

Apr-95 20.531 17.188 20.438 324892800 16.8188

May-95 22.375 19.719 21.172 333260000 17.997

Jun-95 23.094 20.438 22.594 317597200 19.547

Jul-95 27.313 22.125 22.625 582714800 20.922

Aug-95 25.188 21.75 23.125 532686400 21.9908

Sep-95 24.469 20.844 22.625 405582400 22.4282

Oct-95 25.844 20.094 25 557034000 23.1938

Nov-95 25.25 21.063 21.781 467730000 23.0312

Dec-95 23.688 21.281 21.938 413403600 22.8938

Jan-96 23.313 19.969 23.125 514154000 22.8938

Feb-96 25.906 23 24.672 409232400 23.3032

Mar-96 26.766 23.656 25.781 424737600 23.4594

Apr-96 28.469 24.906 28.313 336570800 24.7658

May-96 30 27.406 29.688 301832000 26.3158

Jun-96 31.469 29.063 30.031 320025600 27.697

Jul-96 30.719 26.875 29.469 544224400 28.6564

Aug-96 31.531 29.344 30.625 285966000 29.6252

Sep-96 34.656 30.156 32.969 324858800 30.5564

Oct-96 34.781 32.719 34.313 335158800 31.4814

Nov-96 39.5 34.125 39.219 405242000 33.319

Dec-96 43.063 37.066 41.313 364755800 35.6878

Jan-97 51.625 40.375 51 391445400 39.7628

Feb-97 51.75 47 48.75 324288200 42.919

Mar-97 50.5 43.813 45.844 369409600 45.2252

Apr-97 61.313 44.875 60.75 476863000 49.5314

May-97 64.531 57.438 62 324458800 53.6688

Jun-97 67.469 59.125 63.188 242547400 56.1064

Jul-97 75.375 61.625 70.688 423486400 60.494

Aug-97 72.313 65.5 66.094 292267600 64.544

Sep-97 70.125 65.313 66.156 288403800 65.6252

Oct-97 69.813 61.75 65 383345200 66.2252

Nov-97 71.125 64.5 70.75 228992600 67.7376

Dec-97 73.313 59 64.625 359306000 66.525

Jan-98 75.063 62.188 74.594 371527000 68.225

Feb-98 86 75.25 84.75 350504600 71.9438

Mar-98 90.938 79.25 89.5 274981900 76.8438

Apr-98 99.125 86.625 90.125 254744300 80.7188

May-98 91 81.875 84.813 273845600 84.7564

Jun-98 108.563 83.125 108.375 308288100 91.5126

Jul-98 119.625 105.375 109.938 307324500 96.5502

Aug-98 113.75 95.75 95.938 337509900 97.8378

Sep-98 114.625 94.5 110.063 298081100 101.8254

=========================================

Oct-98 110.125 87.75 105.875 410762300 106.0378

So here is the solution. We will use the 5-month moving average of the Microsoft data to predict Intel. Here is the data:

9.5126

9.7294

9.8918

9.6376

9.398

...

84.7564

91.5126

96.5502

97.8378

101.8254

80.2004

78.8254

77.7754

76.4004

77.3878

The first part of the data is the Microsoft data, the last 5 numbers are the Intel data. Click "Real/+ Exponential", and we have Possibility Confidence*Probability

79.0219 26752

76.242 2048

---------------------------------------------------

Weighted Average

78.8242

Highest Probability

79.0219 26752

Error of each number

1.38994

The most likely outcomes are 79.0219 ± 1. This is a direct hit on the target of 80.9378.

Click "Real/-- Exponential", and we have the similar results:

Possibility Confidence*Probability

79.0219 34576

76.242 7936

81.8018 20184

84.5816 380

87.3615 256

62.3426 64

---------------------------------------------------

Weighted Average

79.6091

Highest Probability

79.0219 34576

Error of each number

1.38994

Customized Software

Customized software can be ordered from Attrasoft upon your request for the following reasons:

Larger External Networks: The current version supports up to 10,000 external neurons. If you want a larger one, it has to be customized. We produce neural networks up to 1,000,000 neurons.
Larger Internal Networks: The current version supports up to 10,000 internal neurons. If you want a larger one, it has to be customized.
Special Data Interface: If you need to interface the Predictor with your application data, a customized software can be ordered. Example, missing data, string data, ...

To get updated information:

http://attrasoft.com

3	DATE	CLOSE
4	Nov-91	375.22
5	Dec-91	417.09
6	Jan-92	408.78
7	Feb-92	412.7
8	Mar-92	403.69
9	Apr-92	414.95
10	May-92	415.35