Creating Custom Data Files for ANNI to Use

Working with custom data files is more technical in nature for someone trying to maximize ANNI's power. These steps can be intimidating to the average user and should only be performed by someone familiar with what they are doing.

  1. First create a regular file for the stock you will train with. This is to give you a base data file to work with.

  2. The data file will be located in a sub-directory called "data" to the directory you saved the ANNI file to. It will be labeled with the stock symbol and .csv (example: AMD would be AMD.csv). Open the data file and view the column and data format used.

  3. ANNI supports up to 5 columns of data (Close, High, Low, Open and Volume). The "Close" column is the target value and must always be available.

  4. Decide which of the remaining columns you will not be using as direct inputs or as inputs to any technical indicators. For example, don't change the Volume column if you use the "Price/Volume Trend" indicator since you will effectively incapacitate this indicator once you change the Volume column.

    Note: The Open values are not used in any indicators so this can be replaced without any adverse effects on indicator calculations.

    IMPORTANT!!
    Replacing the volume column will disrupt the Trade System if any Trade System volume constraints are used. Usage of volume constraints are the default setting in the Trade System. So if you plan to train the Trade System and use it's outputs, either create a separate ANNI file for this purpose or disable the Trade System's volume constraints before proceeding with Trade System Training.

  5. Once you've decided which columns will be available to you, using a spreadsheet program, copy and paste your custom data into the appropriate column and save the file. Examples of custom data can be an Index, Interest Rates, Custom Indicator created by you, another stock's closing values, etc.

    IMPORTANT!!
    ANNI does incremental updates, so it shouldn't overwrite existing column data whenever you do an online update. But, every time you do an update, you must remodify your custom columns new entries with the appropriate data prior to continuing training or you will contaminate the Training Systems with incorrect information.

    IMPORTANT!!
    All custom values must be positive numbers. No negative numbers should exist. If necessary, offset your data by an amount necessary to make them all positive.

  6. If replacing columns other then the Open column, go to ANNI's Prediction System's Options navigation branch and select the Indicators view. Unselect any conflicting indicators as described in step 4, then go to the Raw Data tab and select the columns which you have modified. Then select the Apply button to save the changes.

  7. At this point, ANNI will now be directly using your custom supplied values in its Prediction System.

Copyright © 1999 - 2012 Oolution Technologies. All rights reserved.