Recognition of Offline Handwritten Odia Numerals Using Support Vector Machine

Diabetes mellitus is a group of metabolic diseases in which blood sugar levels are too high. About 8.8% of the world was diabetic in 2017. It is projected that this will reach nearly 10% by 2045. The major challenge is that when machine learning-based classifiers are applied to such data sets for risk stratification, leads to lower performance. Thus, our objective is to develop an optimized and robust machine learning (ML) system under the assumption that missing values or outliers if replaced by a median configuration will yield higher risk stratification accuracy. This ML-based risk stratification is designed, optimized and evaluated, where: (i) the features are extracted and optimized from the six feature selection techniques (random forest, logistic regression, mutual information, principal component analysis, analysis of variance, and Fisher discriminant ratio) and combined with ten different types of classifiers (linear discriminant analysis, quadratic discriminant analysis, naïve Bayes, Gaussian process classification, support vector machine, artificial neural network, Adaboost, logistic regression, decision tree, and random forest) under the hypothesis that both missing values and outliers when replaced by computed medians will improve the risk stratification accuracy. Pima Indian diabetic dataset (768 patients: 268 diabetic and 500 controls) was used. Our results demonstrate that on replacing the missing values and outliers by group median and median values, respectively and further using the combination of random forest feature selection and random forest classification technique yields an accuracy, sensitivity, specificity, positive predictive value, negative predictive value and area under the curve as: 92.26%, 95.96%, 79.72%, 91.14%, 91.20%, and 0.93, respectively. This is an improvement of 10% over previously developed techniques published in literature. The system was validated for its stability and reliability. RF-based model showed the best performance when outliers are replaced by median values.

The combined application of principal component analysis and decision tree in nuclear pulse shape discrimination

In this study, a chemometric approach was established for classifying honey samples based on sugar content, phenolics, antioxidant properties, and color values. Fifty-two different honey samples were obtained from different regions of Turkey, including cotton, chestnut, sunflower, honeydew, citrus, and canola. Sugar content, phenolic profile, total phenolic content, CIE-color values (L*, a*, b*), oxygen radical absorbance capacity, and Trolox equivalent absorbance capacity of honeys were determined. Different display and pattern recognition techniques were used to classify those honey samples. Fifteen variables were reduced to six principal components by principal component analysis, and 78.31% of total variance could be explained. Three main and six sub-clusters were identified in cluster analysis. The model and cross-validation rates for classification by the discriminant analysis (DA) were found as 87.8% and 49.0%, respectively. Artificial neural networks showed better performance than DA, probably due to the model optimization using hidden layers with varying neurons and input variables. Novelty Impact Statement The composition of honey, such as phenolic constituents and sugars, has an impact on its botanical and geographical origins. The physicochemical properties of some Turkish honeys were determined to establish an alternative way to classify their geographical origin by applying chemometric methods using principal component, discriminant and cluster analysis, and artificial neural networks.

BackgroundPiriform aperture is an anatomical region that has been very little studied in terms of sex estimation. Ensemble learning is similarly an unstudied area in sex estimation from human skeletal remains. In this study, it was aimed to perform sex estimation by using the anthropometric measurements of piriform aperture obtained by computed tomography and 3D reconstruction techniques, discriminant function analysis, machine learning algorithms, and ensemble learning method. A total of 442 cases, 226 male and 216 female, aged between 21 and 89 were included in the study. After sex estimation was performed using discriminant analysis, K-nearest neighbor, Gaussian Naive Bayes, multilayer perceptron neural networks, decision trees, support vector machines, and random forest algorithms, a random forest model that accepted the results of these seven methods as predictors was created, and sex estimation was performed again with ensemble learning.ResultsSex prediction results were obtained with a maximum accuracy of 76.5% with discriminant analysis, 84.2% with machine learning algorithms, and 85.7% with the ensemble learning method.ConclusionsIn conclusion, it was seen that piriform aperture showed moderate sexual dimorphism. Sex estimation results could be further improved with machine learning algorithms and especially with the ensemble learning method.

The problems of adulteration and mislabeling are very common in the food industry. Laser-induced breakdown spectroscopy (LIBS) coupled with chemometric methods has many intrinsic advantages on adulteration analysis of various materials. In this work, several chemometric algorithms, i.e., principal component analysis (PCA), decision tree (DT), random forest (RF), partial least squares discriminant analysis (PLS-DA), linear discriminant analysis (LDA), and support vector machine (SVM), were carried out assisted by LIBS technology to study the classification performances on rice geographic origins. A series of samples, including 20 kinds of rice samples from different geographic origins, was detected using LIBS with no pretreatment processes. For data analysis, PCA was employed to reduce the input variables, and to reduce the collinearity of LIBS spectral results as well. The results showed the classification accuracies of the mentioned chemometric algorithms of DT, RF, PLS-DA, LDA, and SVM with 89 input variables of 86.80%, 96.30%, 96.80%, 98.60%, and 99.20%, respectively. At the same time, the operation times of these algorithms were 3.81 s, 54.64 s, 3.63 s, 2.09 s, and 531.01 s, respectively. On the other hand, 30 principal components of input variables were also tested under the same conditions. The classification accuracies for the above algorithms were 81.60%, 98.00%, 95.70%, 98.40%, and 99.20%, respectively. The operation times were 2.01 s, 4.88 s, 3.67 s, 0.36 s, and 308.55 s, respectively. In addition, the five-fold cross-validation classification accuracies with 30 input variables for DT, RF, PLS-DA, LDA, and SVM were 83.75%, 97.95%, 94.75%, 98.35%, and 99.25%, respectively. As a result, LDA was demonstrated to be the best and most efficient tool for rice geographic origin classification assisted by LIBS with high accuracy and analytical speed, which has great potential for rapid identification of adulterated products in agriculture without use of any chemical reagent.

Five cranial nonmetric traits for sex estimation for sex estimation are classified by score according to geometry. The population of origin is one of the factors influencing cranial nonmetric traits. Moreover, among the five cranial traits, the robust traits for estimating sex varied across population. The aim of this study is to suggest the most useful method for sex estimation and demonstrate the need of a suitable method for each population. One-hundred thirty-five three-dimensional skull images from 21st century Korean autopsy cadavers were evaluated using the ordinal scoring system of five cranial nonmetric traits as outlined in Buikstra & Ubelaker (1994). All scores of each trait were analyzed by linear discriminant and decision tree analyses for sex estimation. The frequency of each trait was analyzed and compared to populations from other studies. The accuracy for both sexes was 88.1% by discriminant analysis and 90.4% by decision tree. The traits with the highest accuracy were the glabella and mastoid process in both discriminant analysis and decision tree. Sex estimation in modern Korean cadavers using the cranial nonmetric method was shown to be highly accurate by both discriminant analysis and decision tree. When comparing the pattern of frequency scores in this study with those of other populations, the pattern of trait scores for estimating sex was different for each population, even among populations in the same Asian region, which suggests the need for methods suited for specific populations.

Ink analysis provides a crucial function in forensic document examination for authentication, forgery detection, and ink dating. Today, black gel inks are widely used in both legal and non-legal documents, however they are difficult to analyse due to their pigment formulations, which undergo subtle chemical changes during ageing. Destructive techniques such as Thin Layer Chromatography (TLC) and High-Performance Liquid Chromatography (HPLC) can discriminate ink but are unsuitable for evidentiary purposes. Non-destructive Attenuated Total Reflectance-Fourier Transform infrared (ATR-FTIR) spectroscopy offers an alternative, yet spectral similarities among black gel inks necessitate advanced computational models to facilitate discrimination, classification and age prediction. In this study, predictive ageing models for black gel inks was developed by integrating ATR-FTIR spectroscopy with machine learning (ML). Thirty black gel inks from 23 brands were analysed. Ink lines made using the black gel ink samples were aged for twelve months under three different environmental conditions, and their infrared (IR) spectral data were recorded monthly over the period of 12 months. For age prediction, four classifiers namely Discriminant Analysis(DA), Support Vector Machine(SVM), k-Nearest Neighbour (kNN), and Decision Tree(DT) were trained on full mid-IR, fingerprint region, and PCA (Principal Component Analysis)-derived datasets. Performances of the classifiers were evaluated using accuracy, precision, recall, F1-score, ROC (Receiver Operating Characteristics), and Area Under the Curve (AUC). For age prediction, DA achieved the best accuracy (81.5%) with PCA features, outperforming SVM (76.1%), kNN (48.8%), and DT (40.2%). ROC-AUC values exceeded 90% across all classes. This study demonstrates that ATR-FTIR spectroscopy integrated with machine learning provides a reliable, non -destructive framework for black gel ink classification and age prediction, addressing limitations of destructive methods and strengthening forensic document analysis

This paper carries on research surrounding the influences produced by dimensionality reduction on machine learning classification effect. Firstly, paper constructs the analysis architecture of data dimension reduction classification, combines the two different unsupervised dimension reduction methods, locally linear embedding (LLE) and principal component analysis (PCA) with the five machine learning classification methods: Gradient Boosting Decision Tree (GBDT), Random Forest, Support Vector Machine (SVM), K-Nearest Neighbor (KNN) and Logistic Regression. And then uses the handwritten digital identification dataset to analyze the classification performance of these five classification methods on different dimension datasets by different dimensionality reduction methods. The analysis shows that using the appropriate dimensionality reduction method for dimensionality reduction classification can effectively improve the classification accuracy; the dimensionality reduction classification effect of non-linear dimensionality reduction method is generally better than the linear dimensionality reduction method; different machine learning classification algorithms have significant differences in the sensitivity of dimensions.

The extent of genetic differentiation among 17 Ethiopian populations (249 individuals) of Phytolacca dodecandra (Endod) sampled along altitudinal gradients that varied from 1600 to 3000 m was investigated using random amplified polymorphic DNA (RAPD). The populations were classified into three altitude groups: lowland (1600–2100 m), central-highland (2101–2500 m) and highland (2500–3000 m). Seventy polymorphic loci scored from 12 RAPD primers, singly or in combination with ecogeographical variables (altitude, longitude, latitude, temperature and rainfall), were used for principal component, discriminant, correlation, and stepwise multiple regression analyses. Principal component analysis (PCA) clearly differentiated lowland and the central-highland populations from those of the highlands independent of their geographical regions. Canonical discriminant analysis separated the lowland plants from those of the highlands with the central-highland plants being intermediate. Classificatory discriminant analysis corrected classification of 92.8% of the 249 plants into their respective three altitude groups. Multiple regression analysis identified a strong association between some RAPDs and altitude, temperature and rainfall, while the variation in most RAPDs was explained by combinations of the different ecogeographical variables. It is hypothesised that the different altitude groups may be (1) chemical and/or physiological ecotypes produced as a result of complex interactions of altitude with climatic and/or edaphic factors, or (2) different in ploidy levels. The significant correlations obtained between population means from some RAPDs and altitude and temperature as well as the strong association of some RAPDs with the ecogeographical variables in the multiple regression analysis suggest that part of the RAPD polymorphism could be adaptive, and responsive to environmental selection.

Groundwater is considered as one of the most important sources for water supply in Iran. The Fasa Plain in Fars Province, Southern Iran is one of the major areas of wheat production using groundwater for irrigation. A large population also uses local groundwater for drinking purposes. Therefore, in this study, this plain was selected to assess the spatial variability of groundwater quality and also to identify main parameters affecting the water quality using multivariate statistical techniques such as Cluster Analysis (CA), Discriminant Analysis (DA), and Principal Component Analysis (PCA). Water quality data was monitored at 22 different wells, for five years (2009-2014) with 10 water quality parameters. By using cluster analysis, the sampling wells were grouped into two clusters with distinct water qualities at different locations. The Lasso Discriminant Analysis (LDA) technique was used to assess the spatial variability of water quality. Based on the results, all of the variables except sodium absorption ratio (SAR) are effective in the LDA model with all variables affording 92.80% correct assignation to discriminate between the clusters from the primary 10 variables. Principal component (PC) analysis and factor analysis reduced the complex data matrix into two main components, accounting for more than 95.93% of the total variance. The first PC contained the parameters of TH, Ca2+, and Mg2+. Therefore, the first dominant factor was hardness. In the second PC, Cl-, SAR, and Na+ were the dominant parameters, which may indicate salinity. The originally acquired factors illustrate natural (existence of geological formations) and anthropogenic (improper disposal of domestic and agricultural wastes) factors which affect the groundwater quality.

A comparison between several classification algorithms with feature extraction on real dataset is presented. Principal Component Analysis (PCA) has been used for feature extraction with different values of the ratio R, evaluated and compared using four different types of classifiers on two real benchmark data sets. Accuracy of the classifiers is influenced by the choice of different values of the ratio R. There is no best value of the ratio R, for different datasets and different classifiers accuracy curves as a function of the number of features used may significantly differ. In our cases feature extraction is especially effective for classification algorithms that do not have any inherent feature selections or feature extraction build in, such as the nearest neighbour methods or some types of neural networks.

Rapid and accurate identification of foodborne pathogenic bacteria is of great importance because they are often responsible for the majority of serious foodborne illnesses. The confocal Raman microspectroscopy (CRM) is a fast and easy-to-use method known for its effectiveness in detecting and identifying microorganisms. This study demonstrates that CRM combined with chemometrics can serve as a rapid, reliable, and efficient method for the detection and identification of foodborne pathogenic bacteria without any laborious pre-treatments. Six important foodborne pathogenic bacteria including S. flexneri, L. monocytogenes, V. cholerae, S. aureus, S. typhimurium, and C. botulinum were investigated with CRM. These pathogenic bacteria can be differentiated based on several characteristic peaks and peak intensity ratio. Principal component analysis (PCA) was used for investigating the difference of various samples and reducing the dimensionality of the dataset. Performances of some classical classifiers were compared for bacterial detection and identification including decision tree (DT), artificial neural network (ANN), and Fisher's discriminant analysis (FDA). Correct recognition ratio (CRR), area under the receiver operating characteristic curve (ROC), cumulative gains, and lift charts were used to evaluate the performance of models. The impact of different pretreatment methods on the models was explored, and pretreatment methods include Savitzky-Golay algorithm smoothing (SG), standard normal variate (SNV), multivariate scatter correction (MSC), and Savitzky-Golay algorithm 1st Derivative (SG 1st Der). In the DT, ANN, and FDA model, FDA is more robust for overfitting problem and offers the highest accuracy. Most pretreatment methods raised the performance of the models except SNV. The results revealed that CRM coupled with chemometrics offers a powerful tool for the discrimination of foodborne pathogenic bacteria.

The trustworthiness of Secondary Users (SUs) can be measured through its past and present trust values with sensing reputations given by the neighboring nodes in Cognitive Radio Network (CRN). Basing on these values, it is cleared that whether the SUs can be utilized the free channels of Primary Users (PUs) or not. In this paper, it has been proposed a model to analyze the trustworthiness of Secondary Users (SUs) in Cognitive Radio Network (CRN) with the help of Machine Learning (ML) approaches. It is desired to achieve more accuracy on the predicted data in the process of calculating trustworthiness and spectrum sensing reputation of SUs in Cognitive Radio Network (CRN). It has been also helped to sense the correct number of malicious users, suspicious users and honest users among the total number of SUs. For the simulation work WEKA software has been used, which is a collection of machine learning algorithms for data mining task. Three different types of classifiers of machine learning approaches have been analyzed in this simulation work such as Naive Bayes, Decision Tree and Bayes Network. From this analysis, it is observed that Decision Tree and Bayes Network are performing better than Navies Bayes in terms of providing high accuracy.

In this paper, bangla handwritten digit recognition based post office automation system has been proposed for Bangladeshi Post offices. Actually, postal automation is not any new concept in developed countries, as they are already practicing this. Though Bangladesh has entered in the digital era, postal systems are not digitalized still now. Proposed system automatically sorts mails according to the post code, thus it can save time and money, reducing the necessity of manual sorting. Envelopes of each mails are scanned and region containing post code information is segmented. Different positional and statistical features including Gray Level Co occurances Matrix (GLCM), Local Binary Pattern (LBP), Histogram of oriented gradients (HOG) feature based bangla handwritten digit recognition system is applied to extract each digit. According to post code, mails are stored in postal zone specific buskets. For recognising purpose, Multi Layer Neural Network is applied. Proposed system gains higher accuracy of 99.71%, thus it outperforms state-of-the-art methods.

In recent years, the prevalence of hand-foot-mouth disease (HFMD) in China and some other countries has caused worldwide concern. Mild cases tend to recover within a week, while severe cases may progress rapidly and tend to have bad outcome. Since there is no vaccine for HFMD and anti-inflammatory treatment is not ideal. In this study, we aimed to establish a valid forecasting model for severe HFMD using common laboratory parameters. Retrospectively, 77 severe HFMD cases from Zhengzhou Children's hospital in the peaking period between years 2013 to 2015 were collected, with 77 mild HFMD cases in the same area. The study recorded common laboratory parameters to assist in establishment of the severe HFMD model. After screening the important variables using Mann-Whitney U test, the study also matched the logistic regression (LR), discriminant analysis (DA), and decision tree (DT) to make a comparison. Compared with that of the mild group, serum levels of WBC, PLT, PCT, MCV, MCH, LCR, SCR, LCC, GLO, CK-MB, K, S100, and B in the severe group were higher (p < 0.05), while MCR, EOR, BASOR, SCC, MCC, EO, BASO, NA, CL, T, Th, and Th/Ts were lower (p < 0.05). Five indicators including MCR, LCC, Th, CK-MB, and CL were screened out by LR and the same for DA, and five variables including EO, LCC, CL, GLO, and MCC screened out by DT. The area under the curve (AUC) of LR, DA, and DT was 0.805, 0.779 and 0.864, respectively. The findings were that common laboratory indexes were effectively used to distinguish the mild HFMD cases and severe HFMD cases by LR, DA, and DT, and DT had the best classification effect with an AUC of 0.864.