A NEW METHOD FOR ANALYZING DOWNLOADED DATA FOR STRATEGIC DECISION

Technology assessment survey is nowadays a specific and scientific subject that any manufacture needs for increasing productivity. This function was initially reserved to experts of the studied field. But the increase of information volume has called for a change. Now, we need specialists of technology assessment survey which know about sophisticated methods to extract strategic information from downloaded data. We will explain how to build strategic information. We present here a new and original method of data analysis. This Factorial Relational Analysis is born after 15 years of IBM France mathematics research center works on qualitative data analysis. The method is based on Relational Analysis. The particularity of this method is to work with sparse matrices and to obtain the best classification without any a priori fixation of number of classes. Relational Analysis is used in other sectors than the analysis of matrices issued from downloaded data. For example it is also used in computational lexicography or in credit scoring or in any domain where classification is concerned. Here we choose to present an example of an application in patent analysis.