A study on content-based classification and retrieval of audio database

Utilizing identical climate indices as predictors for all climate divisions within large basins may result in unreliable rainfall forecasts at the sub-basin scale. This study aimed to develop a new approach to identify the most effective predictors among large-scale climate indices for seasonal rainfall forecasting in small areas. The proposed approach combines a selective singular value decomposition method (SSVD) with a non-linear sequential forward selection method (NLSFS). Applying the new algorithm for seasonal rainfall forecasting within two climate divisions in Karkheh basin, Iran, indicated that the climate indices identified by the SSVD differed between the study areas. The combination of these indices exhibited a correlation with seasonal rainfall approximately 11% higher than those derived from the SVD method. Moreover, NLSFS significantly enhanced the forecast accuracy compared to the frequently employed linear sequential forward selection (SFS) method, and the optimal predictors chosen by the two methods differed across all seasons.

Introduction/ Background Cancer is a disease that affects millions worldwide and accurate determination of whether lymph nodes (LNs) near the primary tumor contain metastatic foci is of critical importance for proper patient management. Histopathological evaluation is the only accepted method to make that determination. However, the current standard of care only examines a single central histological section per LN and yields an unacceptable false-negative rate. Aims To help pathologists in their examination we propose a method that extracts textural features from histopathological LN whole slide images (WSI) and then applies support vector machines (SVMs) to automatically identify regions suspicious for metastatic foci. Methods The database consisted of WSI from 44 LNs. Sections were stained with hematoxylin-eosin and examined at 20x (0.45μm resolution). Twenty-eight of the LNs were identified by an expert pathologist as positive for cancer (P), and the remaining sixteen were negative (N). This database was divided into two groups. Group 1 (15P and 5N) was used for training and Group 2 (13P and 11N) was used for testing the classification technique. For all analysis each WSI was divided into non-overlapping 1000 x 1000 pixel sub-images that will be referred to as high-power fields (HPFs). For each LN in Group 1, at least one WSI was annotated by a pathologist to identify rectangular, HPF-scale regions as locally cancerous or locally non-cancerous. From these annotated slides, 924 HPFs (462 P and 462 N) were obtained. For each of these HPFs, statistical features based on gray-level co-occurrence matrices [1] and Law’s texture energy measures [2, 3] were extracted from 9 derived images [4]. The extracted features were submitted to a sequential forward selection (SFS) method [5] to select few non-redundant features providing best class separation (cancerous vs. non-cancerous region). Combinations of the selected features were tested on the 924 HPFs using k-fold cross-validation to find those that produced the best results and consequently to train our SVM-based classifier. In Group 2, WSI were not annotated for cancerous and non-cancerous zones on a HPF scale. Each LN, however, had been labeled by a pathologist as positive or negative for cancer. For each WSI, each section was divided into contiguous HPFs, and those which mainly contain fatty tissue, background, and tears were automatically excluded. Each selected HPFs was classified as cancerous or non-cancerous using the previously trained classifier to obtain the total number of cancer-classified per LN. A receiver operating characteristics (ROC) curve was traced by changing the discriminator threshold (T) used to label the LN as P for cancer as a function of the total number of cancer-classified HPFs. Results During training, 5 Laws features were selected by SFS. Highly satisfactory k-fold cross-validation with a F-score of 0.996 ± 0.005 was obtained using only 2 statistical features computed at different scales. The ROC curve obtained by applying the SVM-classifier to the test set is shown in the next figure. Two valuable operating points can be identified which both guaranteed no false-negative. At T=11 we got 2 false-positives and an optimal F-score of 0.917, and with a more conservative approach, T=1, we got 7 false-positives and a F-score of 0.759. The top-left part of the slide displayed in next figure would have been proposed to the pathologist as the most suspicious region of the cancerous LN.

Sentence alignment using P-NNT and GMM

In this paper, three groups of characteristics related to mass texture are adopted, namely, SGLD (spatial gray level dependence), TS (texture spectrum) and TFCM (texture feature coding method) to describe the characteristics of masses and normal textures on digitized mammograms. Next, under the testing by classifiers, three feature selection methods - SBS (sequential backward selection), SFS (sequential forward selection) and SFSM (sequential floating search method) are used to find out suboptimal subset from 19 features in order to improve the performance of mass detection. Finally, two classifiers PNN (probabilistic neural network) and SVM (support vector machine) are applied and their performances are compared

The inspection of rice grain that may be infected by seedborne disease is important for ensuring uniform plant stands in production fields as well as preventing proliferation of some seedborne diseases. The goal of this study was to use a hyperspectral imaging (HSI) technique to find optimal wavelengths and develop a model for detecting discolored, diseased rice seed infected by bacterial panicle blight (Burkholderia glumae), a seedborne pathogen. For this purpose, the HSI data spanning the visible/near-infrared wavelength region between 400 and 1000 nm were collected for 500 sound and discolored rice seeds. For selecting optimal wavelengths to use for detecting diseased seed, a sequential forward selection (SFS) method combined with various spectral pretreatments was employed. To evaluate performance based on optimal wavelengths, support vector machine (SVM) and linear and quadratic discriminant analysis (LDA and QDA) models were developed for detection of discolored seeds. As a result, the violet and red regions of the visible spectrum were selected as key wavelengths reflecting the characteristics of the discolored rice seeds. When using only two or only three selected wavelengths, all of the classification methods achieved high classification accuracies over 90% for both the calibration and validation sample sets. The results of the study showed that only two to three wavelengths are needed to differentiate between discolored, diseased and sound rice, instead of using the entire HSI wavelength regions. This demonstrates the feasibility of developing a low cost multispectral imaging technology based on these selected wavelengths for non-destructive and high-throughput screening of diseased rice seed.

The paper proposes a method for diagnosing Parkinson’s disease using a reduced set of patients’ voice features samples. The Sequential Forward Selection and Sequential Backward Selection methods were used to reduce features. The data reduced were classified separately with the use of over a dozen popular classifiers. The effectiveness of diagnosing Parkinson’s disease on a reduced set of data was determined for each classifier. Then it was compared with the results obtained for the data without any reduction of features.

In the field of Human-Computer Interaction (HCI), speech emotion recognition technology plays an important role. Facing a small number of speech emotion data, a novel speech emotion recognition method based on feature construction and ensemble learning is proposed in this paper. Firstly, the acoustic features are extracted from the speech signal and combined to form different original feature sets. Secondly, based on Light Gradient Boosting Machine (LightGBM) and Sequential Forward Selection (SFS) method, a novel feature selection method named L-SFS is proposed. And then, the softmax regression model is used to learn automatically the weights of the four single weak learners including Support Vector Machine (SVM), K-Nearest Neighbor (KNN), Extreme Gradient Boosting (XGBoost) and LightGBM. Lastly, based on the learned automatically weights and the weighted average probability voting strategy, an ensemble classification model named Sklex is constructed, which integrates the above four single weak learners. In conclusion, the method reflects the effectiveness of feature construction and the superiority and stability of ensemble learning, and gets good speech emotion recognition accuracy.

Portable Document Format (PDF) files as well as files in several other formats such as (.docx, .hwp and .jpg) are often used to conduct cyber attacks. According to VirusTotal, PDF ranks fourth among document files that are frequently used to spread malware in 2020. Malware detection is challenging partly because of its ability to stay hidden and adapt its own code and thus requiring new smarter methods to detect. Therefore, outdated detection and classification methods become less effective. Nowadays, one of such methods that can be used to detect PDF files infected with malware is a machine learning approach. In this research, the Support Vector Machine (SVM) algorithm was used to detect PDF malware because of its ability to process non-linear data, and in some studies, SVM produces the best accuracy. In the process, the file was converted into byte format and then presented in Byte Frequency Distribution (BFD). To reduce the dimensions of the features, the Sequential Forward Selection (SFS) method was used. After the features are selected, the next stage is SVM to train the model. The performance obtained using the proposed method was quite good, as evidenced by the accuracy obtained in this study, which was 99.11% with an F1 score of 99.65%. The contributions of this research are new approaches to detect PDF malware which is using BFD and SVM algorithm, and using SFS to perform feature selection with the purpose of improving model performance. To this end, this proposed system can be an alternative to detect PDF malware.

Evapotranspiration (ET) has been the focus of research for decades as it is the key to improving water resources management. This study addresses the estimation of daily reference evapotranspiration (ET₀) using machine learning (ML) models with limited meteorological data in newly developed agricultural areas in Egypt. The study tested five methods for selecting key input features to compare the performance of 20 ML models across 14 stations using available daily meteorological data from 1943 to 2024. The modified FAO Penman-Monteith (PM) method was used to estimate ET₀ as a benchmark for evaluating the performance of ML models. The Categorical Boosting (CatBoost) regressor achieved the highest performance across all stations when using the complete feature dataset and maintained high performance with three features. Furthermore, the results showed consistent success of CatBoost, Light Gradient Boosting Machine (LightGBM), and Gradient Boosting Machine (GB) when two features, selected by Sequential Forward Selection (SFS) method, were used. This highlights the feature selection method's ability to capture the most important meteorological variables, making it a valuable tool for improving ET₀ prediction in different agroclimatic regions. In addition, wind speed (U) and maximum temperature (Tmax) were identified as critical predictors of daily ET₀ in Egypt. This approach offers a practical solution for improving water management in newly reclaimed desert areas, not only in Egypt but also in other regions with similar climatic conditions.

In this paper, ant colony optimization (ACO) is applied to hyperspectral band selection. The objective is to select a small band subset such that classification accuracy can be maintained or even improved. The ACO-based band selection technique in this research is independent of any classifier, resulting in lower computational cost. Both supervised (i.e., Jeffries-Matusita distance) and unsupervised (i.e., simplex volume) selection criteria are investigated. The experimental results demonstrate that the classification accuracy on the selected bands is higher than using all bands, and ACO-based methods can outperform the widely used sequential forward selection (SFS) method.

This paper compares kernel-based probabilistic neural networks for speaker verification based on 138 speakers of the YOHO corpus. Experimental evaluations using probabilistic decision-based neural networks (PDBNNs), Gaussian mixture models (GMMs) and elliptical basis function networks (EBFNs) as speaker models were conducted. The original training algorithm of PDBNNs was also modified to make PDBNNs appropriate for speaker verification. Results show that the equal error rate obtained by PDBNNs and GMMs is less than that of EBFNs (0.33% vs. 0.48%), suggesting that GMM- and PDBNN-based speaker models outperform the EBFN ones. This work also finds that the globally supervised learning of PDBNNs is able to find decision thresholds that not only maintain the false acceptance rates to a low level but also reduce their variation, whereas the ad-hoc threshold-determination approach used by the EBFNs and GMMs causes a large variation in the error rates. This property makes the performance of PDBNN-based systems more predictable.

Background and Aim: Therapeutic drug monitoring (TDM) has evolved over the years as an important tool for personalized medicine. Nevertheless, some limitations are associated with traditional TDM. Emerging data-driven model forecasting [e.g., through machine learning (ML)-based approaches] has been used for individualized therapy. This study proposes an interpretable stacking-based ML framework to predict concentrations in real time after olanzapine (OLZ) treatment. Methods: The TDM-OLZ dataset, consisting of 2,142 OLZ measurements and 472 features, was formed by collecting electronic health records during the TDM of 927 patients who had received OLZ treatment. We compared the performance of ML algorithms by using 10-fold cross-validation and the mean absolute error (MAE). The optimal subset of features was analyzed by a random forest-based sequential forward feature selection method in the context of the top five heterogeneous regressors as base models to develop a stacked ensemble regressor, which was then optimized via the grid search method. Its predictions were explained by using local interpretable model-agnostic explanations (LIME) and partial dependence plots (PDPs). Results: A state-of-the-art stacking ensemble learning framework that integrates optimized extra trees, XGBoost, random forest, bagging, and gradient-boosting regressors was developed for nine selected features [i.e., daily dose (OLZ), gender_male, age, valproic acid_yes, ALT, K, BW, MONO#, and time of blood sampling after first administration]. It outperformed other base regressors that were considered, with an MAE of 0.064, R-square value of 0.5355, mean squared error of 0.0089, mean relative error of 13%, and ideal rate (the percentages of predicted TDM within ± 30% of actual TDM) of 63.40%. Predictions at the individual level were illustrated by LIME plots, whereas the global interpretation of associations between features and outcomes was illustrated by PDPs. Conclusion: This study highlights the feasibility of the real-time estimation of drug concentrations by using stacking-based ML strategies without losing interpretability, thus facilitating model-informed precision dosing.

Sequential Forward Selection is one of the most popular variable selection methods and is widely adopted in many applications, among which the design of classifiers. The main problem related to this approach regards the stability of the selection, namely the exploitation of different training data set can lead to the selection of different variable subsets. This problem is particularly critical in applications where variable selection is used to interpret the behaviour of the process or phenomenon under consideration, i.e. to understand which among a potentially huge list of variables actually affect the classification. The paper proposes a method that improves the stability of the sequential forward selection procedure while preserving and possibly improving the classification performance.

Identification, Segmentation, and Image Property Study of Acute Infarcts in Diffusion-Weighted Images by Using a Probabilistic Neural Network and Adaptive Gaussian Mixture Model

In this paper, a new approach for classification of brain tissues into White Matter, Gray Matter, Cerebral Spinal Fluid, Glial Matter, Connective and MS lesion in multiple sclerosis is introduced. This work considers fuzzy multiwavelets, Gaussian Mixture Model (GMM) and Weighted Probabilistic Neural Networks (WPNN) for the classification of the brain tissues. Multiwavelet packet transformation is employed on brain MR images. Since multiwavelet packet transformation yields larger number of subbands compared to multiwavelet and wavelet transformations, we have proposed a fuzzy-set based theory for selection of the subbands. In contrast to the standard method of subband selection, guided by the criteria of signal energy, our method is based on the discriminatory features from the multiwavelet packet transformation coefficients. Singular values are then computed from the selected subbands. The singular values of lower magnitudes are truncated for effective classification of brain tissues in the presence of noise. Probability density functions of the remaining singular values are modeled as GMM. Model parameters are estimated using stochastic EM (SEM). They are used as features for the classification. The classification is carried out using WPNN. Experiments have been carried out using the data sets composed of three modalities of brain MR images, namely T1 and T2 relaxation times and proton density weighted MR images. Experimental results prove that the proposed approach gives better classification rate at various noise levels compared to existing approaches.