Multivariate outlier detection and remediation in geochemical databases

Introduction: Saffron is one of the most important economic plants in the Khorasan province. Awareness of soil quality in agricultural lands is essential for the best management of lands and for obtaining maximum economic benefit. In general, plant growth is a function of environmental factors especially chemical and physical properties of soil (20). It has been demonstrated that there was a positive and high correlation between soil organic matter and saffron yield. Increasing the yield of saffron due to organic matter is probably due to soil nutrient, especially phosphorous and nitrogen and also improvement of soil physical quality (6, 28, 29). The yield of saffron in soils with high nitrogen as a result of vegetative growth is high (8). Shahandeh (6) found that most of the variation of saffron yield depends on soil properties. Due to the economic importance of saffron and the role of soil properties on saffron yield, this research was conducted to find the relationship between saffron yield and some soil physical and chemical properties, and to determine the contribution of soil properties that have the greatest impact on saffron yield in the Ghayenat area. Materials and Methods: This research was performed in 30 saffron fields (30 soil samples) of the Ghayenat area (longitude 59° 10΄ 10.37˝ - 59° 11΄ 38.41˝ and latitude 33° 43΄ 35.08˝ - 33΄ 44΄ 02.78˝), which is located in the Khrasan province of Iran. In this research, 21 soil properties were regarded as the total data set (TDS). Then the principal component analysis (PCA) was used to determine the most important soil properties affecting saffron yield as a minimum data set (MDS) and the stepwise regression to estimate saffron yield. To estimate the yield of saffron in stepwise regression method, saffron yield was considered as a dependent variable and soil physical and chemical properties were considered to be independent variables. Results and Discussion: According to the PCA method, among the 21 studied properties, 7 out of them including calcium, iron, zinc contents, sand, calcium carbonate equivalent percent, mean weight diameter of aggregates (MWD) and manganese (Mn) had the higher Eigenvalues. Therefore, the above properties were introduced as the most important soil properties in saffron fields. Calcium carbonate had the negative effect on the availability of micro-nutrients (26). Christensen et al. (15) found that by increasing the calcium carbonate in soil due to high pH and formation of insoluble components, the uptake of micro-nutrients is especially limited. The results of stepwise regression method (equation 1) showed that soil acidity (pH), zinc content, bulk density, MWD, iron content, salinity (EC), organic carbon and available potassium in soil were the most important properties that affect the yield of saffron, so that the determination coefficient (R2) of the regression model was high (Table 2) and it can explain 74% of the variation of saffron yield. Y = 6924.51 – 1187.31 pH – 89.65 EC + 71.6 Fe – 826.02 Zn + 471.55 OC, + 5490.96 K + 1353.56 BD + 752.82 MWD (1) where Y: saffron yield (kgha-1), pH: soil acidity, EC: electoral conductivity (dSm-1), Fe: iron concentration (mgkg-1), Zn: zinc concentration (mgkg-1), OC: organic carbon (%), K: soil potassium (%), BD: soil bulk density (Mgm-3), and MWD: mean weight diameter of aggregates (MM). Based on the absolute values of standard s in the regression model (Table 3), pH value and then after Zn concentration had the most effect on saffron yield. In general, responses of different plants to soil pH is varied, and saffron grows satisfactory in pH = 7.8 (5). Soil pH influences the uptake of soil nutrients by plants (15), so that this parameter had the most effect on saffron yield and by increasing the soil pH, the yield of saffron decreases. According to the regression model, Zn concentration was the second parameter in saffron yield. Zn has the important role in structure of plant enzymes (30). After these 2 parameters, Bd, MWD, Fe concentration, EC, Organic carbon and K concentration in soil had more effect on saffron yield (Table 3). Conclusion: According to both PCA and regression methods, the concentration of iron and zinc and MWD were determined as the important and effective soil properties on saffron yield in the Ghayenat area. In addition, soil pH in stepwise regression method and calcium carbonate in PCA method were determined as the effective properties on saffron yield. Therefore, it is suggested that the parameters of Zn, Fe, and MWD along with soil pH and calcium carbonate which were regarded individually in two methods, were considered as the most soil properties in saffron yield.

In this work, a real time digital protection algorithm based on PCA and neural network methods is presented for induction motors. The proposed protection algorithm covers internal winding faults, broken rotor bar faults, and bearing faults. Many laboratory experiments have been performed on a specially designed induction motor to evaluate the performance of the suggested protection algorithm. The hybrid protection algorithm uses the instantaneous phase currents. These currents are first preprocessed by PCA to extract distinctive features called residuals. Then the calculated residuals are applied to a feed-forward backpropagation neural network as input vectors. The outputs of the network are winding fault, bearing fault, and normal operating. The proposed algorithm is implemented by using C++ with a NI-DAQ data acquisition board. (5 pages)

Sensitivity study, the premise of the analysis of reservoir damage mechanism, is significant to optimize all the links of drilling and development process, as well as to develop the systematic oil and gas reservoir protection technology scheme. To study a variety of methods developed in recent years for the reservoir sensitivity prediction, it is found that using the combination of single correlation analysis and multiple regression analysis is a new, ideal, and rapid method for the prediction of reservoir sensitivity. Extract the relevant information of sensitivity on the basis of conventional core analysis and sensitivity analysis of minerals, and apply the single correlation analysis and multiple regression analysis to predict the reservoir sensitivity, the accuracy rate can reach 85%, which basically meets the needs of reservoir sensitivity prediction. Compared with the single prediction method, the prediction results obtaining from the combination of the two methods are obviously improved. It is simple, highly applicable, and has a clear physical meaning. The most common method of reservoir sensitivity prediction is sensitivity experiment. Conventional reservoir sensitivity experiment method requires a large number of cores, and the test is a long term procedure, while, during the process of drilling and cutting cores as well as pretreatment for the experiment, different degrees of damage may have been caused to cores. Therefore, conventional evaluation method needs plenty of manpower and material resources, it cannot make sure that the predicting outcome of each time is exactly correct however. Various methods for the fast prediction of reservoir sensitivity have been developed in recent years, such as Elman neural network, BP neural network, grey theory appraising method, multiple discriminate analysis, fuzzy mathematical method and so on, while, there exists defect in each method. Elman network can be regarded as forward neural network which possesses partial memory unit and feedback connection. Its major structure is feedback connection, which is used to memorize the output value of the previous moment, and its convergence speed is slow due to the fixed connection weight. It usually adopts BP neural network model in reservoir sensitivity prediction. However, there exists disadvantages like low convergence speed and failure in global optimization in the conventional BT algorithm towards BP algorithm. Neural network method changes the characteristics of all problems into figures, inference into numerical calculation, while neural network fails to carry out work when the data is insufficient. Besides, it requires corresponding complicated computer program when taking use of neural network method, thus the widespread use of this method has been limited. Grey theory is a new approach aims at studying the uncertainty of less data and poor information, which owns the advantage of being required with less data and small amount of calculation, thus it is extensively used in some certain industries. Within the grey model, the plane nipped between the upper and lower bounds of the forecasted future value is called gray plane, which is unfold in a shape of trumpet, that is to say, the farther in the future time, the larger grey interval of prediction value, in other words, the bigger the grey scale, the smaller the practical significance of the prediction value. Multiple discriminate analysis first divides sensitivity and the known samples of fabric parameter into several groups based on the size of sensitivity, respectively establishing the discriminant functional equation of each group of sample. It is impossible to grouping in detail, thus the predicting outcome is only an interval, but not a specific numerical value. 2nd Workshop on Advanced Research and Technology in Industry Applications (WARTIA 2016) © 2016. The authors Published by Atlantis Press 556 Based on the current method of fast prediction for reservoir sensitivity, this article combines with the previous research results, synthesizes the applied single correlation analysis and multiple regressions analysis to forecast reservoir sensitivity, getting over the disadvantages of other methods in forecasting reservoir sensitivity, supplementing the existed defects. 1. SINGLE CORRECTION ANALYSIS When one or several interrelated variables are taken certain values, although the precise value of its another corresponding variable cannot be confirmed, it still changes within a certain rage in accordance with certain rules. The dependency relation objectively exists among undemanding and uncertain quantities of such phenomenon is known as correlativity. Back in 1980, British statistician Pearson had put forward a calculation formula of measuring the linear relationship between two variables, see formula (1), r usually acts as product moment correlation coefficient. The value range of correlation coefficient r is from -1 to 1; the size of |r| reveals the strong and weak point of the linear relationship between variable x and y. The correlation coefficient value significance and correlative degree evaluation standard see table 1 and table 2.

The remote sensing satellite cluster system, as an important component of the next generation of space architecture in the United States, has important application prospects in the military field. In order to improve the effects of time, with regard to the effectiveness evaluation of the remote sensing satellite cluster system, neural network methods are generally used to satisfy the requirements of real-time decision-making assistance in the military field. However, there are two problems that emerge when applying the existing neural network methods to an effectiveness evaluation of the remote sensing satellite cluster. On the one hand, the neural network model architecture needs to be designed specifically for the remote sensing satellite cluster system. On the other hand, there is still a lack of hyperparameter optimization methods that consume less time and have good optimization effects for the established neural network model. In this regard, two main modifications were made to the back-propagation neural network, to which an effectiveness evaluation was applied. The first comprised a new architecture named BPS, which was designed for the back-propagation neural network so as to improve its prediction accuracy. In BP architecture, one back-propagation neural network is established for each indicator involved in the effectiveness evaluation indicator system of the remote sensing satellite cluster; the output of each back-propagation neural network model is modified to the residual value between the corresponding indicator value and the value that is predicted through a multiple linear regression analysis of the corresponding indicator. The second modification involved the multi-round traversal method, which is based on the three-way decision theory, and it was proposed in order to significantly improve the model’s training time, which is a new type of hyperparameter optimization method. The results show that compared with the traditional simulation model, the modified back-propagation neural network model based on three-way decision theory can quickly and effectively provide stable and accurate evaluation results; this can assist with and meet the requirements for real-time decision-making in the military field.

The Sixth International Symposium on Neural Networks (ISNN 2009)

Protection of an induction motor (IM) against possible faults, such as a stator winding fault, due to thermal deterioration, rotor bar and bearing failures, is very important in environments in which it is used intensively, as in industry as an actuator. In this work, a real time digital protection algorithm based on principal component analysis (PCA) and neural network method is presented for induction motors. The proposed protection algorithm covers internal winding faults (also known as stator faults), broken rotor bar faults, and bearing faults. Many laboratory experiments have been performed on a specially designed induction motor to evaluate the performance of the suggested protection algorithm. The hybrid protection algorithm described in this paper uses is based on a three phase rms supply. These currents are first preprocessed by PCA to extract distinctive features called residuals. The calculated residuals are applied to a feed-forward back-propagation neural network as input vectors for decision making. Outputs of the network are signals denoting winding fault, rotor bar fault, bearing fault, and normal operation. The proposed algorithm is implemented by using Matlab^{TM} and C++ with a NI-DAQ data acquisition board.

A novel online structure damage identification using Principal Component Analysis (PCA) techniques and the perceptron backpropagation neural network is presented. There are three phases to execute this method. In Phase I, system modal information, frequencies and mode shapes, are calculated. Phase II is for damage location identification; the Residual Force Vectors (RFVs) are computed as input to the first neural network. Then the network was trained to simulate damage location identification. Phase III is the severity identification step. The PCA method is used to modify the input for the second neural network. Then this network identifies the severity. There are three advantages of using the PCA method, First, PCA method characterizes the original modal information precisely. Second, PCA method creates the unique data for different damage cases unlike other modal property based data. Third, the accuracy of the damage identification improves significantly, when compared with previously developed methods. This method can be operated online for the real time structural damage identification.

This paper presents a neural network method to locate common fault exactly in a distribution power system (DPS) with renewable sources. The back propagation (BP) neural network method is applied to identify patterns of voltage and current measured from distribution branches. The input matrix of BP network consists of the voltage and current values, which can identify the accurate fault position. The fault location of a common short-circuit fault is analyzed thoroughly in an active distribution network (ADN) with the renewable power sources. Simulation results prove the feasibility and usefulness of the fault location method based on the BP neural network, wherein the fault location accuracy can reach 0.09%.

ObjectiveConsidering the complex and nonlinear relationship between characteristic parameters of tumors (location and morphology) and lung dose‐volume parameters of patients with esophageal cancer (EPC) undergoing intensity‐modulated radiation therapy (IMRT), the model identification method of artificial neural network (ANN) was adopted to build the model to predict lung dose‐volume parameters under different characteristic parameters of tumors. The goal is to enhance the efficiency and quality of radiotherapy planning and reduce the risk of radiation‐induced lung injury (RILI) by providing pre‐planning dose predictions.Methods1). In the previous research work, a retrospective analysis was done on 103 cases of patients with EPC who were treated by radiotherapy, in which a linear regression model was employed and had proved the association between tumor morphology [including tumor relative length (L) and “tumor axial cross‐sectional area” (S)] and lung dose‐volume parameters (including V5, V10, V20 and V30 of lungs). 2). Building on earlier research, this study uses curve regression analysis to further explore the effect of tumor relative position (P) on lung dose‐volume parameters of patients with EPC undergoing IMRT. 3). A backpropagation (BP) neural network model was constructed with P, L, and S as inputs and lung dose‐volume parameters as outputs to establish their quantitative relationship. 4). Genetic algorithm (GA) was combined with BP neural network to optimize the initial weights and thresholds of BP neural network and avoid the model from falling into local optima, so as to improve the performance of BP neural network model.Results1). Quadratic regression models were applied to analyze the relationship of P with V5 and V10 of the lung, yielding adjusted R2 values of 0.177 and 0.081, respectively, with statistically significant regression coefficients (p < 0.01). The results suggest that as P increases, V5 and V10 initially increase, then decrease. Linear regression models were applied to analyze the relationship of P with V20 and V30 of the lung, yielding adjusted R2 values of 0.06 and 0.072, respectively, with significant regression coefficients (p < 0.01). The results show that both V20 and V30 decrease as P increases. 2). The finding of this study indicates that the modeling method of BP neural network can realize the prediction of lung dose‐volume parameters under different characteristic parameters of tumors. The GA‐BP model showed slight improvements in prediction accuracy (PA) and error index (EI) compared to the BP model alone. For both models, prediction accuracy is highest for V5, followed by V10 and V20, with V30 being the least accurate. Notably, both the BP and GA‐BP neural networks achieve outstanding recognition accuracy for V5, with EI and PA values of 0.121 and 88.61% for BP, and 0.113 and 89.00% for GA‐BP, respectively. 3). The re‐optimization of four test cases with under‐predicted whole‐lung V5 and V10 values resulted in substantial improvements over the original plans. Conclusion: The ANN model developed in this study can effectively predict lung dose‐volume parameters for patients with EPC undergoing IMRT. It provides pre‐planning guidance that enhances the efficiency and quality of radiotherapy planning and reducing the risk of RILI.

The incidence and mortality rate of the primary liver cancer are very high and its postoperative metastasis and recurrence have become important factors to the prognosis of patients. Circulating tumor cells (CTC), as a new tumor marker, play important roles in the early diagnosis and individualized treatment. This paper presents an effective method to distinguish liver cancer based on the cellular scattering spectrum, which is a non-fluorescence technique based on the fiber confocal microscopic spectrometer. Combining the principal component analysis (PCA) with back propagation (BP) neural network were utilized to establish an automatic recognition model for backscatter spectrum of the liver cancer cells from blood cell. PCA was applied to reduce the dimension of the scattering spectral data which obtained by the fiber confocal microscopic spectrometer. After dimensionality reduction by PCA, a neural network pattern recognition model with 2 input layer nodes, 11 hidden layer nodes, 3 output nodes was established. We trained the network with 66 samples and also tested it. Results showed that the recognition rate of the three types of cells is more than 90%, the relative standard deviation is only 2.36%. The experimental results showed that the fiber confocal microscopic spectrometer combining with the algorithm of PCA and BP neural network can automatically identify the liver cancer cell from the blood cells. This will provide a better tool for investigating the metastasis of liver cancers in vivo, the biology metabolic characteristics of liver cancers and drug transportation. Additionally, it is obviously referential in practical application.

In this work, an attempt has been made to enhance the diagnostic relevance of spirometric pulmonary function test using neural networks and Principal Component Analysis (PCA). For this study, flow-volume curves (N = 175) using spirometers were generated under standard recording protocol. A method based on neural network is used to predict the most significant parameter, FEV(1). Further, PCA is used to analyze the interdependency of the parameters in the measured and predicted datasets. Results show that the back propagation neural network is able to predict FEV(1) both in normal and abnormal cases. The variation in the magnitude and direction of parameters in the contribution of the principal components shows that FEV(1) is a significant discriminator of normal and abnormal datasets and is further confirmed by the percentage variance in the first few principal components. It appears that this method of prediction and principal component analysis on the measured and predicted datasets could be useful for spirometric pulmonary function test with incomplete data.

In order to improve the efficiency and accuracy of thermal and moisture comfort prediction of underwear, a new prediction model is designed by using principal component analysis method to reduce the dimension of related variables and eliminate the multi-collinearity relationship between variables, and then inputting the converted variables into genetic algorithm (GA) and BP neural network. In order to avoid the problems of slow convergence speed and easy falling into local minimum of Back Propagation (BP) neural network, this paper adopted GA to optimize the weights and thresholds of BP neural network, and utilized MATLAB software to program, and established the prediction models of BP neural network and GA–BP neural network. To verify the superiority of the model, the predicted result of GA–BP, PCA–BP and BP are compared with GA–BP neural network. The results show that PCA could improve the accuracy and adaptability of GA–BP neural network for thermal and moisture comfort prediction. PCA–GA–BP model is obviously superior to GA–BP, PCA–BP, BP, SVM and K-means prediction models, which could accurately predict thermal and moisture comfort of underwear. The model has better accuracy prediction and simpler structure.

Comparison of the Kriging and neural network methods for modeling foF2 maps over North China region

Palmprint recognition system has been one promising biometric system used in Presence System. There are some methods to recognize the individual palmprint as well as to extract its feature. In this research, two recognition methods are compared, i.e., backpropagation neural network and similarity measure using Euclidean distance. While, for feature extraction, we implemented Principal Components Analysis (PCA) method. From the research, it can be concluded that from test results, the best recognition using backpropogation neural networks is 93.33% which is reached when parameters used are: 100 principal components, 1 hidden layer, and 75 neurons. While, implementation of similarity measure using Euclidean distance, the best recognition rate is 96.67% which is reached when 75 principal components are used. When considering the time consumed in recognition, the Euclidean distance gives the better result, i.e. 17.09 seconds, while using backpropagation neural network with 75 neurons, time consumed is 425 seconds. Therefore, from this research, recognition implementation combining both PCA and Euclidean distance are more suggested rather than using combination of PCA and backpropagation neural network.

This study investigated the rapid identification of ceramics via laser-induced breakdown spectroscopy (LIBS) to realize the identification of ancient ceramics from different regions. Ceramics from different regions may have large differences in their elemental composition. Thus, using LIBS technology for ceramic identification is feasible. The spectral intensities of 11 common elements, namely, Si, Al, Fe, Ca, Mg, Ti, Mn, Na, K, Sr, and Ba, in ceramics were selected as classification indices. Principal component analysis (PCA) and kernel principal component analysis (KPCA) combined with the back propagation (BP) neural network were used to identify ceramics. Furthermore, the effects of the PCA and KPCA data processing methods were compared. Finally, this work aimed to select a suitable method for obtaining spectral data on ceramics identified by LIBS through experiments. Results revealed that LIBS technology could aid the routine, rapid, and on-site analysis of archeological objects to rapidly identify or screen various types of objects.