Application of data approximation and classification in measurement systems - comparison of “neural network” and “Least Squares” approximation

Abstract

In measurement systems, environmental conditions are measured based on predefined scenarios. Measured data are then processed in either a decentralized or centralized manner. In advanced systems (especially for distributed data processing), taking artificial intelligence features into consideration could improve measurement performance and reliability. It is assumed as autonomy in measurement system which leads to distributed "intelligent data measurement and processing". In this paper, two different methodologies for "temperature prediction" are compared. A discussion concerning the classification of recorded data is then presented. Both a mathematical approach, the so-called "least squares" approach, and a model-free approach, called back-propagation, are applied and compared for temperature approximation. After approximation, the predicted temperature values are compared with real temperature records for classification purposes. The "classification mechanism" includes signal processing features for improving performance.

Application of Neurocomputing for Data Approximation and Classification in Wireless Sensor Networks

“…In (9), the current and previous values of parameters S ACP , S 1 and S 2 are recorded and used to find the approximation coefficients ( ) that are calculated by solving the over-determined set [ 40 ]. With the new values of approximation parameters ( S ACP , S 1 and S 2 ) and the updated approximation coefficients ( K vector), a new value of S 3 (the value of the “under approximation sensor node”) is calculated.…”

Advanced Bio-inspired Plausibility Checking in a Wireless Sensor Network Using Neuro-immune Systems: Autonomous Fault Diagnosis in an Intelligent Transportation System

“…An optimized neural network was implemented on a wireless sensor network to approximate and classify the records for plausibility checking [4]. The advantages and disadvantages were highlighted in comparison with the classical approaches (like Least Squares) [5] [6]. Artificial immune system features, which are established on human biological immune system to detect and eliminate the threats, could be integrated with different techniques like neural networks to increase the flexibility and accuracy of data processing [7].…”

“…shows the proposed Neuro-immune architecture; to design the network, different parameters have been explored (such as number of layers and neurons) in order to obtain the Maximum data approximation accuracy when the temperature of a sensor is approximated using three neighboring sensor nodes at each cluster[4][6]. Two hidden layers are selected each including four neurons ).…”

Adaptive data sensing rate in ad-hoc sensor networks for autonomous transport application

“…It requires a certain number of input sets to train the network; the number of input sets, the accuracy of the training, and the parameters of the network greatly influence the accuracy of the approximation. In the typical backpropagation technique, the network is trained according to a defined network architecture, number and dimension of input patterns and target vector [22,23,24]. After the training phase, the network is used to approximate data.…”