The MLRP-method for Analysis of Some Problems in Climate and Seismology

258 10 The MLRP-method for Analysis of Some Problems in Climate and Seismology 10.1 Introduction The problem of detection correlations and regularities by data, which is presented by time series or data table, is used in different scientific domains. We have so problem in sociology, medicine, geology and seismology, economics and climate, etc. [Box and Dgenkins, 1974]. The problem of detection correlations and regularities by data is the inverse and ill‐defined problem; therefore, we have to consider every possible information about studied object or process in order to raise solution stability. This information first of all is system of characteristic of studied object: multidimensional and heterogeneity characteristic's space; "noisy" and "duplicate" characteristics; insufficient or excess and missing sample data. We have the most difficult problem, when the priory information about process or object is absent. In addition to that several attendant problems are appeared. Therefore developing new method, we have to consider all above. Firstly, it is necessary to define a class of decision functions (models). Secondary, we must pick out a method of plotting optimal decision function, in other words to define optimality criterion by sample. In the third place, we must to test our model on adequacy and effectiveness (capacity for general conclusion or statistical stability). It is very important to lean against practical and scientific experience of researchers when solving out practical problems. Experts in the certain area own the big share of the aprioristic information and have wide experience in interpretation of results of decisions of practical problems. It is very important in working out and testing of methods. Here it is important to underline an urgency of different countries scientists association at the decision of the important common to all humanity problems, for example, the risk estimation of occurrence the extreme situations due to man‐caused and natural accidents. Unfortunately, availability of data occasionally is the exclusive property of the country or separate business. Therefore, association of scientists and experts of the different countries in work on the general projects is simply common to all humankind necessity. Method presented here depends on statistical information about studied object and can be used in many domains of data mining also. That approach is relied to intelligent analysis to a greater extent, since it uses expert and apriory information generally. At present time, there is joint initiative of the European Commission and European Space Agency, which aims at achieving …