2. Model and Theoretical Fundament

The proposed method considers an automatic multi-variable data analysis of time series obtained from EMN on monitoring points A, B and C geographically distributed in dominant winds direction and bigger population concentration (see Figure 1). The problem is to determine the correlation among all the variables involved in the decision making exercise on health risk for the population. Each monitoring point is considered like a sample point build with different sensors and also with its own perception field. Therefore, each point showed can be seen as a sensors fusion of where the time: series are obtained. In problem solution a self-organized method that uses a Self-Organized Neuronal Network (SOM) has been proposed in order to build an automatic noise suppression method. Neural Networks (NNs) are computational structures and they can learn from examples [^[3]]. In some multi-dimensional engineering problems (like air pollution) is necessary to recognize certain patterns without the necessity of knowing of data nature or their statistical distribution. Some patterns recognition techniques apply NNs to solve problems without the necessity of a prior data distribution knowledge or to make statistical suppositions. Other techniques in pattern recognition have the necessity to make statistical assumptions about data nature like Bayes theorem [^[4], ^[5]]. Consequently, NNs is an ideal tool to solve the problem here exposed due to their operation which is analyzed like a black box that minimizes the energy function [^[6], ^[7]].

2.1. Variable definition

Variables definition is considered like normalized concentration values about pollutants and normalized meteorological values (Wind Speed, Temperature and Relative Humidity). In Table 2, variables are defined in order to build a feature vector x_j and to define a pattern set X_* = {x₁, x₂,.., x_j, .., x_n}. Let X_SO2] be a Sulfur Dioxide set and let X_PM10 be a particles concentration set and their corresponding pattern is defined as x_j = {x₁, x₂, x₄}.