Fuzzy Query FAQ

The purpose of this FAQ is to provide technical answers to the most frequently asked questions about Fuzzy Query. Immediately below you will find several questions; click on the question that is most similar to your question to jump to the section of this document that contains the answer.

• Can I weight the elements of a fuzzy query so that criteria A has more influence on the QCIX than criteria B?
  
• After I create a Fuzzy Query file (*.fqf), I would like to associate it with a different data source than the one I created it with. Can I do this?
  
• After I create a Fuzzy Query file I made some changes to the associated data source, and now the query does not run. Why?
  
• Can I create a unique fuzzy set shape, or am I limited to the pre-defined shapes available on the Fuzzy Sets tab? 
  
• I have been using Fuzzy Query 1.0 and now have version 1.1 but I can't seem to open my 1.0 files. What can I do?

Can I weight the elements of a fuzzy query so that criteria A has more influence on the QCIX than criteria B?

Yes. Suppose you have a stock market database with columns for Price/Earnings, Growth, and Capitalization, and you have defined a low, medium and high fuzzy set for the P/E and Growth columns, as well as a small, mid, large, and huge fuzzy set for the Capitalization column. Further suppose that you have defined the following Fuzzy Criteria: 

• A: Price/Earnings is/are low 
• B: Growth is/are high 
• C: Capitalization is/are small 
• D: Capitalization is/are mid 

To give a criteria more weight in an AND statement simply, repeat the statement multiple times. Note: you must be using the Average AND processing option in the Query Controls screen. Given the above example, you could give Growth twice the weight of Price/Earnings by forming the expression: (A & B & B). If you only wanted Growth to be 50% more important than P/E, you could accomplish this with
the expression: (A & A & B & B & B). Here Growth has a relative weighting of 3 (since there are 3 Bs) and P/E has a relative weighting of 2 (since there are 2 As).

Let's examine an expression that produces the following weighted criteria: 

To get the appropriately weighted fuzzy criteria simply repeat the symbols in the AND
expression the number of times equal to their relative weights. Hence we have the expression: ( A & A & A & B & B & (C | D) )

In this example we are interested in both small and mid cap stocks. Since a stock may be both small and mid cap at the same time, we want to capture the higher of the two membership values so we use an OR expression (C | D). We include that expression once, which gives us the weighting we want. 

[Top of Page]

After I create a Fuzzy Query file (*.fqf), I would like to associate it with a different data source than the one I created it with. Can I do this? 

Unfortunately Not. Fuzzy Query Versions 1.0 and 1.1 require that the *.fqf and data source files be tightly bound together. Look for this to be changed in the next major update to Fuzzy Query. 

[Top of Page]

After I create a Fuzzy Query file I made some changes to the associated data source, and now the query does not run. Why? 

In Fuzzy Query Versions 1.0 and 1.1, the *.fqf and data source files are tightly bound together. You can add and delete rows from a data source, but you can not change an existing column definition, add a column, or delete a column. Look for this to be changed in the next major update to Fuzzy Query. 

[Top of Page]

Can I create a unique fuzzy set shape, or am I limited to the pre-defined shapes available on the Fuzzy Sets tab? 

Fuzzy set definitions are very important, and should represent the semantic meaning of the user. Pre-defined fuzzy set shapes are provided for quick definition of fuzzy sets, and the Edit Fuzzy Set dialog box is provided to allow the user to change the shape of a defined fuzzy set. One of the Edit options is a multi-point fuzzy set shape. Using this option, the user can create a uniquely shaped fuzzy set to better represent the userís intent and meaning. 

I have been using Fuzzy Query 1.0 and now have version 1.1 but I can't seem to open my 1.0 files. What can I do? 

Due to a change in the underlying architecture of our Fuzzy System Engine, we were unable to maintain backward compatibility with the version 1.0 files. However, this will not be a problem in future releases of Fuzzy Query.