Improve Classification Performance Research Articles

Unified Text-Image Space Alignment with Cross-Modal Prompting in CLIP for UDA

Unsupervised Domain Adaptation (UDA) aims to transfer models trained on a labeled source domain to an unlabeled target domain. Due to the excellent generalization ability of Vision Language Models (VLMs) such as CLIP in downstream tasks, most recent methods apply CLIP to UDA tasks through learning domain-specific text prompts for source and target domains separately. However, these methods fail to dynamically adjust image features based on the characteristics of their respective domains, thereby limiting their alignment with domain-specific text prompts in CLIP’s joint space, which is a key factor in improving classification performance in the target domain. To bridge this gap, we propose a Unified Text-Image Space Alignment with Cross-Modal Prompting (UTISA) framework for UDA. Firstly, we introduce a cross-modal prompt learning (CMP) module to generate domain-specific image prompts and layer-specific image prompts for the visual branch to encode domain-specific knowledge globally and locally. Secondly, under the guidance of image prompts, we introduce a domain-aware multi-layer feature fusion (DMF) module to construct multi-layer domain features for each domain and enhance the image features with these multi-layer domain features, which enables the image features to better reflect the characteristics of their respective domains, thereby promoting their alignment with domain-specific text prompts. Moreover, we introduce a perturbation-driven regularization (PDR) mechanism for the target domain to enhance the robustness and generalization of the model. The experiments demonstrate that UTISA achieves the best performance on three mainstream UDA benchmarks, including 87.9% on OfficeHome, 90.9% on VisDA-2017, 62.4% on DomainNet.

Diagnostic classification in toxicologic pathology using attention-guided weak supervision and whole slide image features: a pilot study in rat livers.

The diagnostic classification of digitized tissue images based on histopathologic lesions present in whole slide images (WSI) is a significant task that eludes modern image classification techniques. Even with advanced methods designed for digital histopathology, the domain of toxicologic pathology presents challenges in that histopathologic features may be at times complex, subtle, and/or rare. We propose an innovative weakly supervised learning method that leverages minimal annotations, a state-of-the-art self-supervised vision transformer for embedding extraction, and a novel guided attention mechanism that is better suited for heavily imbalanced datasets typical in toxicologic pathology. Our model demonstrates improvements in diagnostic classification and attention heatmap quality over the previously described clustering-constrained-attention multiple-instance learning method on several lesion classes in rat livers (38% improvement in AUC). We also demonstrate how an ensemble of binary classifiers improves interpretability and allows for multiclass classification and the classification of diagnostic regions of interest in each slide. The improved classification performance and higher contrast heatmaps better support toxicologic pathologists' histopathology analysis and will enable more efficient workflows as they are further refined and integrated into routine use.

Optimizing combination of feature extraction and classifiers in supervised classification of cells images

Developments in the medical field have opened the opportunity to conduct analyses on a personalized patient level. One of the important analyses that can be conducted is the cellular response to engineered materials, and the most appropriate non-invasive methods are imaging. These images of the cells are unstained brightfield images, as they are acquired from multiparametric microfluidic chambers in the presence of biomaterials and fluids that can change the optical path length over time as the cells’ health state is monitored. These experimental conditions lead to an image dataset with unique illumination, texture, and noise spectrum. This study explores the optimization of supervised cell classification by combining feature extraction architectures and machine learning classifiers, with a focus on applications in biomaterial risk assessment. Brightfield microscopy images of three cell types (A549, BALB 3T3, and THP1) were analyzed to evaluate the impact of Inception V3, Squeeze Net, and VGG16 architectures paired with classifiers including KNN, Decision Tree, Random Forest, AdaBoost, Neural Networks, and Naïve Bayes. Dimensionality reduction using Information Gain was applied to improve computational efficiency and accuracy. Butterworth filters with varying parameters were used to balance the enhancement of image features and noise removal, improving classification performance in certain cases. Experimental results demonstrate that the VGG16 architecture, when paired with Neural Networks, achieves higher classification accuracy as measured by different metrics. The improved accuracy when using Butterworth filters compared to the unfiltered dataset and the differences between various Butterworth filters indicate the importance of optimizing filter parameters for these types of images.

Training State-of-the-Art Deep Learning Algorithms with Visible and Extended Near-Infrared Multispectral Images of Skin Lesions for the Improvement of Skin Cancer Diagnosis

An estimated 60,000 people die annually from skin cancer, predominantly melanoma. The diagnosis of skin lesions primarily relies on visual inspection, but around half of lesions pose diagnostic challenges, often necessitating a biopsy. Non-invasive detection methods like Computer-Aided Diagnosis (CAD) using Deep Learning (DL) are becoming more prominent. This study focuses on the use of multispectral (MS) imaging to improve skin lesion classification of DL models. We trained two convolutional neural networks (CNNs)—a simple CNN with six two-dimensional (2D) convolutional layers and a custom VGG-16 model with three-dimensional (3D) convolutional layers—using a dataset of MS images. The dataset included spectral cubes from 327 nevi, 112 melanomas, and 70 basal cell carcinomas (BCCs). We compared the performance of the CNNs trained with full spectral cubes versus using only three spectral bands closest to RGB wavelengths. The custom VGG-16 model achieved a classification accuracy of 71% with full spectral cubes and 45% with RGB-simulated images. The simple CNN achieved an accuracy of 83% with full spectral cubes and 36% with RGB-simulated images, demonstrating the added value of spectral information. These results confirm that MS imaging provides complementary information beyond traditional RGB images, contributing to improved classification performance. Although the dataset size remains a limitation, the findings indicate that MS imaging has significant potential for enhancing skin lesion diagnosis, paving the way for further advancements as larger datasets become available.

Algorithms for data imputation based on entropy

Recent advancements in data imputation have focused on various machine learning techniques, including methods like mean, median, and mode imputation, along with more complex approaches like k-nearest neighbors (KNN) and multiple imputation by chained equations (MICE). Research into entropy-based methods offers a promising direction. This method minimizes uncertainty by selecting imputation values that reduce the overall entropy of the dataset. The goal of this work is to develop an algorithm that imputes missing data by minimiz-ing conditional entropy, thus ensuring that the missing values are filled in a way that pre-serves the relationships between the variables. The method is designed for both qualitative and quantitative data, including discrete and continuous variables, aiming to reduce uncer-tainty in classification tasks and enhance the performance of machine learning models. The proposed algorithm is based on conditional entropy minimization, using entropy as a measure of uncertainty in data. For each incomplete row, the algorithm computes the con-ditional entropy for possible imputation values. The value that minimizes conditional entropy is selected, as it reduces uncertainty in the target variable. This process is iterated for each missing value until all missing data is imputed. Three types of tests were performed on two datasets. The analysis showed that the pro-posed algorithms are quite slow compared to other methods and can be improved, for exam-ple, by multiprocessing, as described in our work [15]. The type 1 test showed that the pro-posed algorithms do not give a gain on the RMS deviation metric, but significantly reduce en-tropy (type 2 test). At the same time, these methods show an improvement in classification performance over the baseline models (type 3 test). Thus, the proposed entropy-based imputation methods have shown good results and can be considered by researchers as an additional tool to improve the accuracy of decision mak-ing, but further computational optimisation studies are needed to improve the performance of these methods. The algorithm shows promise in improving classification accuracy by selecting imputa-tion values that minimize conditional entropy. Future research will focus on optimizing the method for large datasets and expanding its application to various domains.

Multi-Skin disease classification using hybrid deep learning model

Among the many cancers that people face today, skin cancer is among the deadliest and most dangerous. As a result, improving patients’ chances of survival requires skin cancer to be identified and classified early. Therefore, it is critical to assist radiologists in detecting skin cancer through the development of Computer Aided Diagnosis (CAD) techniques. The diagnostic procedure currently makes heavy use of Deep Learning (DL) techniques for disease identification. In addition, skin lesion extraction and improved classification performance are achieved through Region Growing (RG) based segmentation. At the outset of this study, noise is reduced using an Adaptive Wiener Filter (AWF), and hair is removed using a Maximum Gradient Intensity (MGI). Then, the best RG, which is the result of integrating RG with the Modified Honey Badger Optimiser (MHBO), does the segmentation. Finally, several forms of skin cancer are classified using the DL model MobileSkinNetV2. The experiments were conducted on the ISIC dataset and the results show that the accuracy and precision were improved to 99.01% and 98.6%, respectively. In comparison to existing models, the experimental results show that the proposed model performs competitively, which is great news for dermatologists treating cancer.

Recognition of autism in subcortical brain volumetric images using autoencoding-based region selection method and Siamese Convolutional Neural Network

BackgroundAutism Spectrum Disorder (ASD) is a neurodevelopmental condition that affects social interactions and behavior. Accurate and early diagnosis of ASD is still challenging even with the improvements in neuroimaging technology and machine learning algorithms. It’s challenging because of the wide range of symptoms, delayed appearance of symptoms, and the subjective nature of diagnosis. In this study, the aim is to enhance ASD recognition by focusing on brain subcortical regions, which are critical for understanding ASD pathology. MethodologyFirst, subcortical structures were extracted from a collection of brain MRI datasets using sophisticated processing steps. Next, a 3D autoencoder was trained on these 3D images to help identify brain regions related to ASD. Two distinct feature selection methods were then applied to the features extracted from the encoder. The highest-ranked features were iteratively selected and increased to reconstruct a specific percentage of the brain that represents the most relevant parts for ASD. Finally, a Siamese Convolutional Neural Network (SCNN) was employed as the classifier model. ResultsThe 3D autoencoder stage helped in identifying and reconstructing the significant subcortical regions related to ASD. Based on the studied dataset, high agreement in regions like the Putamen and Pallidum indicated the critical nature of these structures in distinguishing Autism from controls cases. Subsequently, applying SCNN on these selected subcortical regions yielded promising results. For example, using the classifier on the output regions identified by the Mutual Information (MI) features selection method achieved the highest accuracy of 0.66. ConclusionsThis study shows that using a two-stage model involving autoencoder and SCNN can notably improve the classification of ASD from brain MRI volumetric images. Applying an iterative feature extraction approach allowed to achieve a more accurate identification of ASD-related brain areas. This two-stage approach not only improved classification performance but also enhanced the interpretability of the neuroimaging data.

Enhanced thyroid disease prediction using ensemble machine learning: a high-accuracy approach with feature selection and class balancing

Thyroid disorders are increasingly prevalent, making early detection crucial for reducing mortality and complications. Accurate prediction of disease progression and understanding the interplay of clinical features are essential for effective diagnosis and treatment. Our study addresses these challenges by employing a standard machine learning model, enhanced with comprehensive clinical feature analysis and an ensemble learning technique. By leveraging machine learning, we can identify key risk factors and improve diagnostic accuracy. To achieve optimal prediction outcomes, we evaluated seventeen machine learning models and implemented an Ensemble ML classifier using a hard voting strategy. Class balancing techniques, particularly random oversampling, significantly improved classification performance. Our experimental results demonstrate that the proposed model outperforms existing methods, achieving 100% sensitivity and 99.72% accuracy using the XGBoost algorithm and SelectKBest feature selection. By addressing feature reduction and high class-imbalance, the ensemble ML classifier with hard voting proves more effective in handling classification challenges.

A comprehensive evaluation of histopathology foundation models for ovarian cancer subtype classification

Histopathology foundation models show great promise across many tasks, but analyses have been limited by arbitrary hyperparameters. We report the most rigorous single-task validation study to date, specifically in the context of ovarian carcinoma morphological subtyping. Attention-based multiple instance learning classifiers were compared using three ImageNet-pretrained encoders and fourteen foundation models, each trained with 1864 whole slide images and validated through hold-out testing and two external validations (the Transcanadian Study and OCEAN Challenge). The best-performing classifier used the H-optimus-0 foundation model, with balanced accuracies of 89%, 97%, and 74%, though UNI achieved similar results at a quarter of the computational cost. Hyperparameter tuning the classifiers improved performance by a median 1.9% balanced accuracy, with many improvements being statistically significant. Foundation models improve classification performance and may allow for clinical utility, with models providing a second opinion in challenging cases and potentially improving the accuracy and efficiency of diagnoses.

SRF: SpectrumRecombineFormer for Hyperspectral Image Classification

Hyperspectral (HS) imaging is a valuable technique for accurately classifying materials because of the abundance of spectral information and high resolution it provides. However, the characteristics of Hyperspectral images (HSI), such as high-dimensional features and information redundancy, pose significant challenges to data processing. Traditional dimensionality reduction methods often have information loss, high computational complexity, and easy to ignore the strong correlation between HSI spectral bands when dealing with HSI data. Although other methods can achieve satisfactory classification performance, they do not consider the dimensionality reduction of HSI, and they focus on the model performance, which limits further improvement in classification performance. This paper proposes a transformer-based framework called “SpectrumRecombineFormer” (SRF), which is composed of two key modules, namely “Spatial Spectral ReCombination” (SSRC) and “Cross-layer Fusion” (CF). The SSRC is capable of utilizing both adjacent and non-adjacent spectrums to generate the spatial-sequential perceptive representations, which alleviates the effect of the strong correlation between HSI spectral bands. The CF can avoid the loss of information during the feed-forward procedure among layers. Extensive experiments on five existing datasets (widely-adopted Indian Pines, Houston2013, Pavia University, Salinas and KSC) demonstrate the capability of our proposed method to address the above mentioned challenges. Both quantitative and qualitative experimental ablation studies, including visualization results, reveal that the proposed SRF method can successfully and efficiently classify hyperspectral images and surpass the other state-of-the-art methods. For access to the source code, please visit https://github.com/kangpeilun/SRF-HSI-Classification-master .

An Explainable Artificial Intelligence Text Classifier for Suicidality Prediction in Youth Crisis Text Line Users: Development and Validation Study.

Suicide represents a critical public health concern, and machine learning (ML) models offer the potential for identifying at-risk individuals. Recent studies using benchmark datasets and real-world social media data have demonstrated the capability of pretrained large language models in predicting suicidal ideation and behaviors (SIB) in speech and text. This study aimed to (1) develop and implement ML methods for predicting SIBs in a real-world crisis helpline dataset, using transformer-based pretrained models as a foundation; (2) evaluate, cross-validate, and benchmark the model against traditional text classification approaches; and (3) train an explainable model to highlight relevant risk-associated features. We analyzed chat protocols from adolescents and young adults (aged 14-25 years) seeking assistance from a German crisis helpline. An ML model was developed using a transformer-based language model architecture with pretrained weights and long short-term memory layers. The model predicted suicidal ideation (SI) and advanced suicidal engagement (ASE), as indicated by composite Columbia-Suicide Severity Rating Scale scores. We compared model performance against a classical word-vector-based ML model. We subsequently computed discrimination, calibration, clinical utility, and explainability information using a Shapley Additive Explanations value-based post hoc estimation model. The dataset comprised 1348 help-seeking encounters (1011 for training and 337 for testing). The transformer-based classifier achieved a macroaveraged area under the curve (AUC) receiver operating characteristic (ROC) of 0.89 (95% CI 0.81-0.91) and an overall accuracy of 0.79 (95% CI 0.73-0.99). This performance surpassed the word-vector-based baseline model (AUC-ROC=0.77, 95% CI 0.64-0.90; accuracy=0.61, 95% CI 0.61-0.80). The transformer model demonstrated excellent prediction for nonsuicidal sessions (AUC-ROC=0.96, 95% CI 0.96-0.99) and good prediction for SI and ASE, with AUC-ROCs of 0.85 (95% CI 0.97-0.86) and 0.87 (95% CI 0.81-0.88), respectively. The Brier Skill Score indicated a 44% improvement in classification performance over the baseline model. The Shapley Additive Explanations model identified language features predictive of SIBs, including self-reference, negation, expressions of low self-esteem, and absolutist language. Neural networks using large language model-based transfer learning can accurately identify SI and ASE. The post hoc explainer model revealed language features associated with SI and ASE. Such models may potentially support clinical decision-making in suicide prevention services. Future research should explore multimodal input features and temporal aspects of suicide risk.

Application of Symbolic Classifiers and Multi-Ensemble Threshold Techniques for Android Malware Detection

Android malware detection using artificial intelligence today is a mandatory tool to prevent cyber attacks. To address this problem in this paper the proposed methodology consists of the application of genetic programming symbolic classifier (GPSC) to obtain symbolic expressions (SEs) that can detect if the android is malware or not. To find the optimal combination of GPSC hyperparameter values the random hyperparameter values search method (RHVS) method and the GPSC were trained using 5-fold cross-validation (5FCV). It should be noted that the initial dataset is highly imbalanced (publicly available dataset). This problem was addressed by applying various preprocessing and oversampling techniques thus creating a huge number of balanced dataset variations and on each dataset variation the GPSC was trained. Since the dataset has many input variables three different approaches were considered: the initial investigation with all input variables, input variables with high feature importance, application of principal component analysis. After the SEs with the highest classification performance were obtained they were used in threshold-based voting ensembles and the threshold values were adjusted to improve classification performance. Multi-TBVE has been developed and using them the robust system for Android malware detection was achieved with the highest accuracy of 0.98 was obtained.

Enhancing Zero-Shot Learning Through Kernelized Visual Prototypes and Similarity Learning

Zero-shot learning (ZSL) holds significant promise for scaling image classification to previously unseen classes by leveraging previously acquired knowledge. However, conventional ZSL methods face challenges such as domain-shift and hubness problems. To address these issues, we propose a novel kernelized similarity learning approach that reduces intraclass similarity while increasing interclass similarity. Specifically, we utilize kernelized ridge regression to learn visual prototypes for unseen classes in the semantic vectors. Furthermore, we introduce kernel polarization and autoencoder structures into the similarity function to enhance discriminative ability and mitigate the hubness and domain-shift problems. Extensive experiments on five benchmark datasets demonstrate that our method outperforms state-of-the-art ZSL and generalized zero-shot learning (GZSL) methods, highlighting its effectiveness in improving classification performance for unseen classes.

Optimizing Reservoir Separability in Liquid State Machines for Spatio-Temporal Classification in Neuromorphic Hardware

In this paper, we propose an optimization approach using Particle Swarm Optimization (PSO) to enhance reservoir separability in Liquid State Machines (LSMs) for spatio-temporal classification in neuromorphic systems. By leveraging PSO, our method fine-tunes reservoir parameters, neuron dynamics, and connectivity patterns, maximizing separability while aligning with the resource constraints typical of neuromorphic hardware. This approach was validated in both software (NEST) and on neuromorphic hardware (SpiNNaker), demonstrating notable results in terms of accuracy and low energy consumption when using SpiNNaker. Specifically, our approach addresses two problems: Frequency Recognition (FR) with five classes and Pattern Recognition (PR) with four, eight, and twelve classes. For instance, in the Mono-objective approach running in NEST, accuracies ranged from 81.09% to 95.52% across the benchmarks under study. The Multi-objective approach outperformed the Mono-objective approach, delivering accuracies ranging from 90.23% to 98.77%, demonstrating its superior scalability for LSM implementations. On the SpiNNaker platform, the mono-objective approach achieved accuracies ranging from 86.20% to 97.70% across the same benchmarks, with the Multi-objective approach further improving accuracies, ranging from 94.42% to 99.52%. These results show that, in addition to slight accuracy improvements, hardware-based implementations offer superior energy efficiency with a lower execution time. For example, SpiNNaker operates at around 1–5 watts per chip, while traditional systems can require 50–100 watts for similar tasks, highlighting the significant energy savings of neuromorphic hardware. These results underscore the scalability and effectiveness of PSO-optimized LSMs on resource-limited neuromorphic platforms, showcasing both improved classification performance and the advantages of energy-efficient processing.

Deep Learning for Risky Cardiovascular and Cerebrovascular Event Prediction in Hypertensive Patients

In this comprehensive study, we employed a versatile approach to tackle the prediction challenges associated with atrial fibrillation (AF) and cardiovascular events (CE). Exploiting the Gaussian copula synthesizer technique for data generation, we created high-quality synthetic data to overcome the limitations posed by scarce patient records. Heart rate variability (HRV), known to be an efficient indicator of cardiac health often used with artificial intelligence (AI), was used to train and optimize custom-built deep learning (DL) models. Additionally, we explored transfer learning (TL) to enhance the model capabilities by adapting our AF classification model to address CE classification challenges, effectively transferring learned features and patterns, without extensive retraining. As a result, our models achieved accuracy rates of 77% for AF and 82% for CEs, with high sensitivity, highlighting the efficacy of synthetic data generation and transfer learning in improving classification performance across diverse medical datasets. These findings hold significant promise for enhancing diagnostic and predictive capabilities in clinical settings, ultimately contributing to improved patient care and outcomes.

Recognition of Human Activities via SSAE Algorithm: Implementing Stacked Sparse Autoencoder

This study evaluates the performance of Stacked Sparse Autoencoder (SSAE) combined with Support Vector Machine (SVM) against a standard SVM for classification tasks. We assessed both models using accuracy, precision, sensitivity, and F1 score. The SSAE Support Vector Machine significantly outperformed the standard SVM, achieving an accuracy of 89% compared to 37%. SSAE also achieved higher precision (87% vs. 75%) and sensitivity (89% vs. 37%), with an F1 score of 88% versus 36% for the standard SVM. These results indicate that SSAE enhances the model’s ability to capture complex patterns and provide reliable predictions. This study highlights the effectiveness of SSAE in improving classification performance, suggesting further research with larger datasets and additional optimization techniques to maximize model efficiency

DCTCNet: Sequency discrete cosine transform convolution network for visual recognition.

The discrete cosine transform (DCT) has been widely used in computer vision tasks due to its ability of high compression ratio and high-quality visual presentation. However, conventional DCT is usually affected by the size of transform region and results in blocking effect. Therefore, eliminating the blocking effects to efficiently serve for vision tasks is significant and challenging. In this paper, we introduce All Phase Sequency DCT (APSeDCT) into convolutional networks to extract multi-frequency information of deep features. Due to the fact that APSeDCT can be equivalent to convolutional operation, we construct corresponding convolution module called APSeDCT Convolution (APSeDCTConv) that has great transferability similar to vanilla convolution. Then we propose an augmented convolutional operator called MultiConv with APSeDCTConv. By replacing the last three bottleneck blocks of ResNet with MultiConv, our approach not only reduces the computational costs and the number of parameters, but also exhibits great performance in classification, object detection and instance segmentation tasks. Extensive experiments show that APSeDCTConv augmentation leads to consistent performance improvements in image classification on ImageNet across various different models and scales, including ResNet, Res2Net and ResNext, and achieving 0.5%-1.1% and 0.4%-0.7% AP performance improvements for object detection and instance segmentation, respectively, on the COCO benchmark compared to the baseline.

Smart Distillation for Land Use Monitoring: A Lightweight Remote Sensing Segmentation Strategy

As the demand for land use monitoring continues to grow, high-precision remote sensing products have become increasingly important. Compared to traditional methods, deep learning networks demonstrate significant advantages in automatic feature extraction, handling complex scenes, and improving classification accuracy. However, as the complexity of these networks increases, so does the computational cost. To address this challenge, we propose an innovative knowledge distillation model, integrating two key modules—spatial-global attention feature distillation (SGAFD) and channel attention-based relational distillation (CARD). This model enables a lightweight “student” network to be guided by a large “teacher” network, enhancing classification performance while maintaining a compact model size. We validated our approach on the large-scale public remote sensing datasets GID15 and LoveDA, and the results show that these modules effectively improve classification performance, overcoming the limitations of lightweight models and advancing the practical applications of land use monitoring.

A Large-Scale Agricultural Land Classification Method Based on Synergistic Integration of Time Series Red-Edge Vegetation Index and Phenological Features.

Agricultural land classification plays a pivotal role in food security and ecological sustainability, yet achieving accurate large-scale mapping remains challenging. This study presents methodological innovations through a multi-level feature enhancement framework that transcends traditional time series analysis. Using Shandong Province, northern China's agricultural heartland, as a case study, we first established a foundation with time series red-edge vegetation indices (REVI) from Sentinel-2 imagery, uniquely combining the normalized difference red edge index (NDRE705) and plant senescence reflectance index (PSRI). Moving beyond conventional time series analysis, we innovatively amplified key temporal characteristics through newly designed spatial feature parameters (SFPs) and phenological feature parameters (PFPs). This strategic enhancement of critical temporal points significantly improved classification performance by capturing subtle spatial patterns and phenological transitions that are often overlooked in traditional approaches. The study yielded three significant findings: (1) The synergistic application of NDRE705 and PSRI significantly outperformed single-index approaches, demonstrating the effectiveness of our dual-index strategy; (2) The integration of SFPs and PFPs with time series REVI markedly enhanced feature discrimination at crucial growth stages, with PFPs showing superior capability in distinguishing agricultural land types through amplified phenological signatures; (3) Our optimal classification scheme (FC6), leveraging both enhanced spatial and phenological features, achieved remarkable accuracy (93.21%) with a Kappa coefficient of 0.9159, representing improvements of 4.83% and 0.0538, respectively, over the baseline approach. This comprehensive framework successfully mapped 120,996 km2 of agricultural land, differentiating winter wheat-summer maize rotation areas (39.44%), single-season crop fields (36.16%), orchards (14.49%), and facility vegetable fields (9.91%). Our approach advances the field by introducing a robust, scalable methodology that not only utilizes the full potential of time series data but also strategically enhances critical temporal features for improved classification accuracy, particularly valuable for regions with complex farming systems and diverse crop patterns.

A hybrid explainable model based on advanced machine learning and deep learning models for classifying brain tumors using MRI images

Brain tumors present a significant global health challenge, and their early detection and accurate classification are crucial for effective treatment strategies. This study presents a novel approach combining a lightweight parallel depthwise separable convolutional neural network (PDSCNN) and a hybrid ridge regression extreme learning machine (RRELM) for accurately classifying four types of brain tumors (glioma, meningioma, no tumor, and pituitary) based on MRI images. The proposed approach enhances the visibility and clarity of tumor features in MRI images by employing contrast-limited adaptive histogram equalization (CLAHE). A lightweight PDSCNN is then employed to extract relevant tumor-specific patterns while minimizing computational complexity. A hybrid RRELM model is proposed, enhancing the traditional ELM for improved classification performance. The proposed framework is compared with various state-of-the-art models in terms of classification accuracy, model parameters, and layer sizes. The proposed framework achieved remarkable average precision, recall, and accuracy values of 99.35%, 99.30%, and 99.22%, respectively, through five-fold cross-validation. The PDSCNN-RRELM outperformed the extreme learning machine model with pseudoinverse (PELM) and exhibited superior performance. The introduction of ridge regression in the ELM framework led to significant enhancements in classification performance model parameters and layer sizes compared to those of the state-of-the-art models. Additionally, the interpretability of the framework was demonstrated using Shapley Additive Explanations (SHAP), providing insights into the decision-making process and increasing confidence in real-world diagnosis.