Fisher Linear Discriminant Method Research Articles

Automatic recognition of effective and interference signals based on machine learning: A case study of acoustic emission and electromagnetic radiation

Integrated monitoring technology using acoustic emission (AE) and electromagnetic radiation (EMR) is a promising mean for monitoring coal and rock dynamic disasters. However, due to the complex underground environment, blasting, drilling and mechanical operations may generate interference signals that affect the accuracy of early warning. It is essential to study the automatic recognition of effective and interference signals for AE and EMR. For this reason, a field test of synchronous AE-EMR monitoring was conducted in a coal mine. Further, the time domain, frequency domain and fractal characteristics of effective and interference signals for AE and EMR were analyzed, and sensitive characteristics were obtained by the Relief algorithm. Based on this, automatic recognition models of effective and interference signals were established by Fisher's linear discriminant method, support vector machine and ensemble learning method, respectively. Field application showed that support vector machine performed higher recognition accuracy. The AE and EMR warning indicators were obtained based on the signal recognition model, and the results show that the indicators reflect the risk of rock burst more accurately and earlier than the original indicators before the removal of interference signals.

Short Words for Writer Identification Using Neural Networks

In biometrics, it is desirable to distinguish a person using only a short sample of his handwriting. This problem is treated in the present work using only a short word with three letters. It is shown that short words can contribute to high-performance writer identification if line characteristics are extracted using morphological directional transformations. Thus, directional morphological structuring elements are used as a tool for extracting this kind of information with the morphological opening operation. The line characteristics are organized based on Markov chains so that the elements of the transition matrix are used as feature vectors for identification. The Markov chains describe the alternation in the directional line features along the word. The analysis of the feature space is carried out using the Fisher linear discriminant method. The identification performance is assessed using neural networks, where the simplest neural structures are sought. The capabilities of these simple neural structures are investigated theoretically concerning the achieved separability into the feature space. The identification capabilities of the neural networks are further assessed using the leave-one-out method. It is proved that the neural methods achieve identification performance that approaches 100%. The significance of the proposed method is that it is the only one in the literature that presents high identification performance using only one short word. Furthermore, the features used as well as the classifiers are simple and robust. The method is independent of the language used regardless of the direction of writing. The NIST database is used for extracting short-length words having only three letters each.

Component Analysis Model of Glass Relics Based on Logistic Regression and Fisher Linear Discrimination

In recent years, with the attention paid by various countries to the exploitation and protection of cultural relics and the continuous development of industrialization, people have gradually realized the necessity of analyzing the weathering of ancient glass relics. In this paper, the physical characteristics, chemical composition and content of sample cultural relics are analyzed and their types are identified by using multiple algorithms such as Logistic regression model, Fisher linear discrimination, Fisher’s exact test, etc. The study of this issue can provide guiding suggestions for the analysis and classification of ancient cultural materials, and help archaeologists carry out a series of work such as cultural relics protection. The results show that the "type" factor has the greatest relationship with weathering, followed by "color" and "decoration". The regression equation between the sample and weathering chemical composition is established, and it is concluded that "type" and "SnO2" have a large negative correlation with weathering, "MgO" and "SO2" have a large positive correlation, and other variables have a weak correlation. Finally, Fisher linear discriminant method is used to identify the category of unknown glass relics. By estimating the training group and testing the test group, it is found that the algorithm can correctly classify 100.0% of the original grouped cases. We can identify the unknown cultural relics and find that A1, A5, A6 and A7 cultural relics belong to high potassium glass, and the rest belong to lead barium glass. In order to test the accuracy of the model in identifying the category of glass relics, Fisher's exact test algorithm was used. It was found that the classification results were more sensitive to silica in the fluctuation range of - 50% to - 25%, and less sensitive to other chemical components, it is stable.

FFV-MBC: A Novel Fused Finger-Vein Recognition Method Based on Monogenic Binary Coding

<p>To improve pattern representation capabilities and robustness in traditional finger-vein recognition algorithms. In this paper, we propose FFV-MBC, a novel fused finger-vein recognition method based on monogenic binary coding (MBC). First of all, the amplitude, orientation, and phase information of the finger-vein images are filtered by a multi-scale monogenic log-Gabor filter and encoded by the binary coding theory. Three local features, MBC-A, MBC-P, and MBC-O, are achieved from different combinations of local image intensity and variation coding. After obtaining the features, we utilize the block-based Fisher Linear Discriminant method to reduce the dimension. Finally, the similarity components are calculated by the cosine distance and fused for the final finger-vein recognition results. We evaluate our proposed method on two publicly available datasets and one self-built dataset, i.e., Malaysian Polytechnic University (FV-USM), the Group of Machine Learning and Applications of Shandong University (SDUMLA-HMT), and our team, Signal and Information Processing Laboratory (FV-SIPL). On average, the proposed method achieved high recognition accuracy, i.e., 99.30%, and 1.10% equal error rates (EER). Overall, the proposed method performs better than most classical and state-of-the-art finger-vein recognition methods.</p> <p> </p>

Classification and identification of glass products based on K-means clustering analysis and Fisher's linear discriminant method

Ancient glass is very susceptible to weathering due to the buried environment, and the internal elements of the weathered glass will exchange with the elements in the environment thus leading to changes in the composition ratio. In order to classify and identify glass objects, this paper analyzes the classification laws of high potassium glass and lead-barium glass based on the data of samples of known types; selects the appropriate chemical composition for each category and then divides it into subcategories, gives specific division methods to establish a classification model, and uses this model for the chemical composition of glass artifacts of unknown categories to identify the types to which they belong

Ultra-High Frequency Partial Discharge Signal Recognition in GIS based on Fisher Linear Discriminant Theory

In this paper, dual tree complex wavelet transform (DT-CWT) is used to decompose UHF partial discharge (PD) signal in multi-scale. The optimal decomposition level of complex wavelet is solved. The wavelet energy features of real part and imaginary part of UHF-PD signal under the optimal decomposition scale are extracted. Fisher linear discriminant method is used to select the features of energy features and to identify the PD types. The results show that the optimized real part and imaginary part high frequency wavelet energy features can effectively identify four kinds of typical insulation defects, and the recognition rate can reach to 92.5%. Meanwhile, the optimal complex wavelet energy (OCWEF) feature has better sensitivity and recognition effect in PD type identification.

Multivariate analysis using a discriminant method for evaluating the techniques of weed management in soybean crop

Abstract Background: The analysis of information generated from experiments involving different treatments, can be done by multivariate statistical analysis techniques, such as discriminant analysis, to analyze data obtained from predefined groups. Objective: Verify, through discriminant analysis, the differences among cover crop (Avena strigosa, Chenopodium quinoa, Cichorium intybus, and fallow land) treatments with respect to main crop soybean yield. Methods: For weed control, these cover crops were subjected to different management techniques, namely mowing, the application of glyphosate or the application of paraquat. The experimental design consisted of completely randomized blocks in a 4 × 3 × 2 factorial scheme, with four replications, consisting of the following factors: Factor A: (treatment) cover crops of A. strigosa, C. quinoa, C. intybus, and fallow land; Factor B: (management) plots were subdivided and treated with the application of paraquat or glyphosate, or the mowing of cover plants; Factor C: the plots were sub-subdivided and managed by one or two applications of a post-emergence herbicide. In order to evaluate the percentage of correct classifications of the different management techniques and treatments, a data matrix was elaborated for evaluation of variables relating to the soybean crop and the data were standardized by log - log 10 - log (n; 10). Multivariate analysis was performed using Fisher's linear discriminant method. Results: Discriminant analysis selected four variables with discriminatory power relating to the A. strigosa, C. quinoa, C. intybus and fallow, which contributed to 100% of the explained variance. Conclusions: Treatment with oats used as a cover crop provided higher soybean crop yield, whereas in terms of management, weed control using glyphosate provided the best results with all cover crops.

Neighborhood multi-granulation rough sets-based attribute reduction using Lebesgue and entropy measures in incomplete neighborhood decision systems

For incomplete data with mixed numerical and symbolic attributes, attribute reduction based on neighborhood multi-granulation rough sets (NMRS) is an important method to improve the classification performance. However, most classical attribute reduction methods can only handle finite sets as to produce more attributes and lower classification accuracy. This paper proposes a novel NMRS-based attribute reduction method using Lebesgue and entropy measures in incomplete neighborhood decision systems. First, some concepts of optimistic and pessimistic NMRS models in incomplete neighborhood decision systems are given, respectively. Then, a Lebesgue measure is combined with NMRS to study neighborhood tolerance class-based uncertainty measures. To analyze the uncertainty, noise and redundancy of incomplete neighborhood decision systems in detail, some neighborhood multi-granulation entropy-based uncertainty measures are developed by integrating Lebesgue and entropy measures. Inspired by both algebraic view with information view in NMRS, the pessimistic neighborhood multi-granulation dependency joint entropy is proposed. What is more, the corresponding properties are further deduced and the relationships among these measures are discussed, which can help to investigate the uncertainty of incomplete neighborhood decision systems. Finally, the Fisher linear discriminant method is used to eliminate irrelevant attributes to significantly reduce computational complexity for high-dimensional datasets, and a heuristic attribute reduction algorithm with complexity analysis is designed to improve classification performance of incomplete and mixed datasets. Experimental results under seven UCI datasets and eight gene expression datasets illustrate that the proposed method is effective to select most relevant attributes with higher classification accuracy, as compared with representative algorithms.

Micro-crack detection method of steel beam surface using stacked autoencoders on massive full-scale sensing strains

Accurate micro-crack detections on the whole surface of civil structures have great significance. Distributed optical fiber sensor based on Brillouin optical time-domain analysis technology exhibits great facility to measure strain distributions along the whole surface of structures with a high spatial resolution, thus providing a potential and competitive solution to the detection problem. However, mainly due to low signal-to-noise ratio in measurements, such sensor system is still limited in crack detection–based structural health monitoring applications. How to extract high-quality micro-crack feature representations from the low signal-to-noise ratio–distributed strain measurements is crucial to solve the problem. It has been demonstrated in field of pattern recognition that deep learning can automatically extract high-quality noise-robust feature representations from mass chaos data. Therefore, a micro-crack detection method is proposed herein based on deep learning to analyze the full-scale strain measurements. Each measurement is normalized and segmented into a set of equal-length subsequences. Autoencoders, a typical kind of building block of deep neural network, are stacked layer-wise into a deep network and then exploited to automatically extract feature representations from the subsequences. Each extracted feature representation is labeled as one of the two categories by a Softmax regression. One category originates in the subsequences acquired from structure sections with crack defects and another from sections without any cracks. The micro-crack detections are achieved by solving such a crack/non-crack binary classification problem. A 15-m-long steel I-beam with artifact crack defects is built up in laboratory to verify the proposed method. Experimental results demonstrate that the minimum size of detectable crack opening width reaches to 23 μm, and besides, the proposed method is significantly better than traditional Fisher linear discriminant analysis method and classical support vector machine on the detection accuracy.

A Gene selection approach based on the fisher linear discriminant and the neighborhood rough set

ABSTRACTIn recent years, tumor classification based on gene expression profiles has drawn great attention, and related research results have been widely applied to the clinical diagnosis of major gene diseases. These studies are of tremendous importance for accurate cancer diagnosis and subtype recognition. However, the microarray data of gene expression profiles have small samples, high dimensionality, large noise and data redundancy. To further improve the classification performance of microarray data, a gene selection approach based on the Fisher linear discriminant (FLD) and the neighborhood rough set (NRS) is proposed. First, the FLD method is employed to reduce the preliminarily genetic data to obtain features with a strong classification ability, which can form a candidate gene subset. Then, neighborhood precision and neighborhood roughness are defined in a neighborhood decision system, and the calculation approaches for neighborhood dependency and the significance of an attribute are given. A reduction model of neighborhood decision systems is presented. Thus, a gene selection algorithm based on FLD and NRS is proposed. Finally, four public gene datasets are used in the simulation experiments. Experimental results under the SVM classifier demonstrate that the proposed algorithm is effective, and it can select a smaller and more well-classified gene subset, as well as obtain better classification performance.

Discriminant Model of Coal Mining Microseismic and Blasting Signals Based on Waveform Characteristics

Generally, there are two important types of microseismic (MS) signals caused by mining and blasting activities at coal mines. The waveform characteristics of MS signals using FFT, STA/LTA method, and envelope analysis were studied to distinguish these two types of MS signals. The main results are as follows: the dominant frequency and duration of two types of signals are significantly different. The following peak envelope curves of two types of MS signals fit a power function. The power exponent was obtained to describe the attenuated speed of the MS signals. The attenuation of the coal mining MS signals is slower and more fluctuant than that of the blasting signal. Waveform characteristics consisting of the dominant frequency, duration, and attenuation coefficient were extracted as the discriminating parameters. The discriminating performance of these parameters was compared and discussed. Based on the waveform characteristics, a discriminant model for coal mining MS and blasting signals was established by using Fisher linear discriminant method and its performance was checked. The results show that the accuracy of the discriminant model is more than 85%, which can meet the requirements of MS monitoring at coal mines.

An Effective Security Based Wireless Information System Based on Fusion Principal Component Analysis

In the real world applications, wireless networks are an integral part of day-to-day life for many people, with businesses and home users relying on them for connectivity and communication. This paper examines the problems relating to the topic of wireless security and the background literature. The biometric systems often face limitations because of sensitivity to noise, intra class invariability, data quality, and other factors. Improving the performance of individual matchers in the aforementioned situation may not be effective. Multi biometric systems are used to overcome this problem by providing multiple pieces of evidence of the same identity. This system provides effective fusion scheme that combines information presented by the multiple domain experts based on the Rank level fusion integration method, thereby increasing the efficiency of the system which is not possible by the unimodal biometric system. The proposed multimodal biometric system has a number of unique qualities, starting from utilizing principal component analysis and fisher's linear discriminant methods for individual matchers authentication and the novel rank level fusion method is used in order to consolidate the results obtained from different biometric matchers. The ranks of the individual matchers are combined using highest rank, Borda count, and logistic regression method. From the results it can be concluded that the overall performances of the wireless security based multi biometric systems are improved even in the presence of quality of data.

Fisher Linear Discriminant Method for Forest Fire Risk Point on Transmission Line

Fisher Linear Discriminant Method for Forest Fire Risk Point on Transmission Line

A Three-Component Fisher-Based Feature Weighting Method for Supervised PolSAR Image Classification

This letter presents a feature weighting method for polarimetric synthetic aperture radar (PolSAR) image classification. Appropriate feature weighting is essential for obtaining accurate classifications but so far has remained an open research problem. We propose in this letter a supervised three-component feature weighting method based on the Fisher linear discriminant. Fisher linear discriminant method is used to calculate a coefficient for each feature. Then, these coefficients are modified according to a three-component scattering power decomposition model, combining both physical and statistical scattering characteristics to adapt them for the particular scattering mechanisms inherent in PolSAR data and assigned to the coherency matrix to enhance the discriminating ability of the features. Freeman decomposition and Wishart classifier are used to classify the PolSAR image. The effectiveness of the proposed method is demonstrated by experiments NASA/JPL AIRSAR L-band and CSA Radarsat-2 C-band PolSAR images of the San Francisco area.

Quantitative Analysis on Robustness of FLD and PCA-based Face Recognition Algorithms

Principal Component Analysis (PCA) has emerged as a more efficient approach for extracting features for many pattern classification problems. It has been the standard approach to reduce the high-dimensional original pattern vector space into low-dimensional feature vector space, that removes some of the noisy directions. PCA is an unsupervised technique which does not include label information of the data. In addition to PCA, another method Fisher Linear Discriminant (FLD) analysis has been widely used. In this paper, we report experimental results to quantify the robustness of PCA and FLD methods for face recognition. The experimentation was performed based on different levels of additive noise and rotations in handling face recognition problem. FLD outperforms the traditional PCA on the basis of robustness.

Variable selection for Fisher linear discriminant analysis using the modified sequential backward selection algorithm for the microarray data

One of the major challenges is small sample size as compared to large features number for microarray data. Variable selection is an important step for improving diagnostics of cancer or the classification according to the phenotypes via gene expression data. In this study, we propose a modified sequential backward selection (SBS) algorithm to deal with the case where the covariance matrix is singular. Then we propose a variable selection algorithm based on the weighted Mahalanobis distance and modified SBS methods. Furthermore, based on the proposed variable selection algorithm, a Fisher linear discriminant method is proposed to improve the accuracy of tumor classification through simultaneously taking into account genes’ joint discriminatory power. To validate the efficiency, we apply the proposed discriminant method to two different DNA microarray data sets for experiment investigation. The empirical results show that our method for tumor classification can obtain better classification effectiveness than Markov random field method and independent variable group analysis I methods, which demonstrates that the proposed variable selection method can obtain more correct and informative gene subset if taking into account the joint discriminatory power of genes for tumor classification.

Multicomposition analysis and pattern recognition of Chinese geographical indication product: vinegar

Multicomposition fingerprints with several chemical compositions containing inorganic elements and organic compounds (amino acids, polyhydric alcohols, organic acids) were measured to distinguish two geographical indication-protected vinegars (GIs) from general vinegars (nGIs). The two of GIs were named Shanxi extra aged vinegar and Zhenjiang vinegar from Shanxi and Jiangsu province, respectively. Principal component analysis and Fisher linear discriminant methods were applied for the pattern recognition and classification of GI product. It was not suitable by simply using one kind of composition to make a distinction between GIs and nGIs. However, by using multicomposition to build a classify model, the classification of SXVs was described by Co, As, Al, Mg, Ca, erythritol, arabitol, sorbitol, proline, lysine and pyruvic acid, while ZJVs classification was described by threonine, serine, glycine, lysine, Ba, erythritol, xylitol and lactic acid. The GI samples can be classified with high accuracy according to the discriminant model, of which the false rate is 3.88 % in SXVs model and 10.85 % in ZJVs model. This method can be a useful method for protecting the geographical indication vinegars from the fake or adulterate vinegar commodities.

A Novel Face Recognition Algorithm

This paper firstly analyzes the principle of face recognition algorithm, studies feature selection and distance criterion problem, puts forward the defects of PCA face recognition algorithm and LDA face recognition algorithm. According to the deficiencies and shortcomings of PCA face recognition algorithm and LDA face recognition algorithm, this paper proposes a solution -- PCA+LDA. The method uses the PCA method to reduce the dimensionality of feature space, it uses Fisher linear discriminant analysis method to classification, the realization of face recognition. Experiments show that, this method can not only improve the feature extraction speed, but also the recognition rate is better than single PCA method and LDA method.

A GA-based feature selection approach with an application to handwritten character recognition

In the framework of handwriting recognition, we present a novel GA–based feature selection algorithm in which feature subsets are evaluated by means of a specifically devised separability index. This index measures statistical properties of the feature subset and does not depends on any specific classification scheme. The proposed index represents an extension of the Fisher Linear Discriminant method and uses covariance matrices for estimating how class probability distributions are spread out in the considered N-dimensional feature space. A key property of our approach is that it does not require any a priori knowledge about the number of features to be used in the feature subset. Experiments have been performed by using three standard databases of handwritten digits and a standard database of handwritten letters, while the solutions found have been tested with different classification methods. The results have been compared with those obtained by using the whole feature set and with those obtained by using standard feature selection algorithms. The comparison outcomes confirmed the effectiveness of our approach.

Bi Modal Person Identification System

Face recognition is one of the most important applications of computer vision in recent years. The developed multimodal biometric system possesses a number of unique qualities, starting from utilizing principal component analysis and Fisher's linear Discriminant methods for individual matcher's identity authentication and utilizes the novel feature fusion method to consolidate the results obtained from different biometric matchers. In this paper, we present a bimodal face-finger recognition system that fuses results from both Principal Component Analysis and Fisher face projections. The proposed approach is tested on a real database consisting of 500 images and shows promising results compared to other techniques. The main goal of bi modal identification system is to develop the security system for the areas that require high level of security. The Receiver Operating Characteristics also shows that the proposed method is superior compared to other techniques under study.