Hello,
I found this in the documentation of IR Expert
Knowledge engineers may tailor the adaptive learning terms that are associated with a solution. New terms may be added or the weight given to currently assigned terms may be adjusted. For instance, in the above example, if the knowledge engineer wants to make sure that DOC 1 is chosen as the solution for any query involving the term blurry or fuzzy, she can then adjust the frequency of those terms in association with that document. By adjusting the frequency, she will be increasing the weight given to the document relative to other possible solutions, which will force the document to the top of the query list.
Can anyone help me with this one, I don't know how to adjust the frequency of the terms (keywords) associated with a solution.
We are working with well defined keywords (and synonyms)linked to every solution (as it was in Tivoli)
I'm searching for a way to give every keyword the same weight. The results we are getting now are sometimes not correct (for us). Ir Expert returns us in some cases not the result we want. Example, when I search for a solution and I have a match on 4 "keywords", it is possible (with the automatic calculation - natural log....) that IR Expert calculates a higher score for solutions with only 2 keywords than the score for solutions with 4 matching keywords.
Can I give the same "weight" to every keyword so that I get a search simply based on the number of corresponding keywords found ?
Is there another way to get IR Export to execute a simple keyword (somewhat like in Tivoli)search without the automatic calculation of IDF (natural log(TDOC/TFREQ)+1)
Thanks !!!
I found this in the documentation of IR Expert
Knowledge engineers may tailor the adaptive learning terms that are associated with a solution. New terms may be added or the weight given to currently assigned terms may be adjusted. For instance, in the above example, if the knowledge engineer wants to make sure that DOC 1 is chosen as the solution for any query involving the term blurry or fuzzy, she can then adjust the frequency of those terms in association with that document. By adjusting the frequency, she will be increasing the weight given to the document relative to other possible solutions, which will force the document to the top of the query list.
Can anyone help me with this one, I don't know how to adjust the frequency of the terms (keywords) associated with a solution.
We are working with well defined keywords (and synonyms)linked to every solution (as it was in Tivoli)
I'm searching for a way to give every keyword the same weight. The results we are getting now are sometimes not correct (for us). Ir Expert returns us in some cases not the result we want. Example, when I search for a solution and I have a match on 4 "keywords", it is possible (with the automatic calculation - natural log....) that IR Expert calculates a higher score for solutions with only 2 keywords than the score for solutions with 4 matching keywords.
Can I give the same "weight" to every keyword so that I get a search simply based on the number of corresponding keywords found ?
Is there another way to get IR Export to execute a simple keyword (somewhat like in Tivoli)search without the automatic calculation of IDF (natural log(TDOC/TFREQ)+1)
Thanks !!!
