Kaled Mohamed Almhdi1,3, Paolo Valigi1, Vidas Gulbinas2, Rainer Westphal3 and Rainer Reuter3
1. Universita' di Perugina, Dipartimento di Ingegneria Elettronica e dell'Informazione,
06125 Perugia, Italy
2. Institute of Physics, 2053 Vilnius, Lithuania
3. Universität Oldenburg, Institut für Physik, 26111 Oldenburg, Germany
Artificial neural networks (ANN) have been involved in many applications to solve real world problems. In commercial purposes ANNs can be applied to predict the profit, market movements, and price level based on the market’s historical dataset. In medical applications, doctors can evaluate the case of many patients depending on the historical dataset of other patients who had the same case. In industry, engineers can apply ANNs to solve many engineering problems such as classifications, prediction, pattern recognition, and non-linear problems where the issues are very difficult or might be impossible to solve through normal mathematical processes. ANNs have been applied to predict slant path rain attenuation (1), to predict rain attenuation on an Earth-space path, to predict water quality index (WQI), and to signal predictions in a nuclear power plant (2). They have also been used in face recognition (3). In medical applications, ANNs have been utilized in detecting brain disease (4) and DNA ploidy, as well as cell cycle distribution of breast cancer aspirate cells that are measured by image cytometry and analysed by ANNs for their prognostic significance (5).
Support vector machines (SVM) are modern and effective tools that have already been examined to solve difficulties such as classification problems and pattern recognition. SVMs can be used to solve more complex problems when compared to ANNs. In SVMs there is no need to select features from several applications, and SVMs have demonstrated that they are more accurate and stable than ANNs, which will be proven for fluorescence spectra classification later in this paper. EARSeL eProceedings x, issue/year 2 SVMs have been applied to medical binary classification problems (6), to recognize radar emitter signals (7), to detect complicated attacks (8), and to visual speech recognition (9), and many other applications.
This paper reports on oil classification with fluorescence spectroscopy. The objective is to classify the oil fluorescence spectra based on a laboratory dataset of fluorescence spectra of several oil classes (sludge, crude and heavy oil). The classification was carried out using the following three methods: channel relationship method (CRM), artificial neural networks (ANNs), and support vector machines (SVMs).
To read this full paper please click the "Download Document" button below: