Proposed Classification Method Research Articles

A mutual reconstruction network model for few-shot classification of histological images: addressing interclass similarity and intraclass diversity.

The automated classification of histological images is crucial for the diagnosis of cancer. The limited availability of well-annotated datasets, especially for rare cancers, poses a significant challenge for deep learning methods due to the small number of relevant images. This has led to the development of few-shot learning approaches, which bear considerable clinical importance, as they are designed to overcome the challenges of data scarcity in deep learning for histological image classification. Traditional methods often ignore the challenges of intraclass diversity and interclass similarities in histological images. To address this, we propose a novel mutual reconstruction network model, aimed at meeting these challenges and improving the few-shot classification performance of histological images. The key to our approach is the extraction of subtle and discriminative features. We introduce a feature enhancement module (FEM) and a mutual reconstruction module to increase differences between classes while reducing variance within classes. First, we extract features of support and query images using a feature extractor. These features are then processed by the FEM, which uses a self-attention mechanism for self-reconstruction of features, enhancing the learning of detailed features. These enhanced features are then input into the mutual reconstruction module. This module uses enhanced support features to reconstruct enhanced query features and vice versa. The classification of query samples is based on weighted calculations of the distances between query features and reconstructed query features and between support features and reconstructed support features. We extensively evaluated our model using a specially created few-shot histological image dataset. The results showed that in a 5-way 10-shot setup, our model achieved an impressive accuracy of 92.09%. This is a 23.59% improvement in accuracy compared to the model-agnostic meta-learning (MAML) method, which does not focus on fine-grained attributes. In the more challenging, 5-way 1-shot setting, our model also performed well, demonstrating a 18.52% improvement over the ProtoNet, which does not address this challenge. Additional ablation studies indicated the effectiveness and complementary nature of each module and confirmed our method's ability to parse small differences between classes and large variations within classes in histological images. These findings strongly support the superiority of our proposed method in the few-shot classification of histological images. The mutual reconstruction network provides outstanding performance in the few-shot classification of histological images, successfully overcoming the challenges of similarities between classes and diversity within classes. This marks a significant advancement in the automated classification of histological images.

When an extra rejection class meets out-of-distribution detection in long-tailed image classification

Detecting Out-of-Distribution (OOD) inputs is essential for reliable deep learning in the open world. However, most existing OOD detection methods have been developed based on training sets that exhibit balanced class distributions, making them susceptible when confronted with training sets following a long-tailed distribution. To alleviate this problem, we propose an effective three-branch training framework, which demonstrates the efficacy of incorporating an extra rejection class along with auxiliary outlier training data for effective OOD detection in long-tailed image classification. In our proposed framework, all outlier training samples are assigned the label of the rejection class. We employ an inlier loss, an outlier loss, and a Tail-class prototype induced Supervised Contrastive Loss (TSCL) to train both the in-distribution classifier and OOD detector within one network. During inference, the OOD detector is constructed using the rejection class. Extensive experimental results demonstrate that the superior OOD detection performance of our proposed method in long-tailed image classification. For example, in the more challenging case where CIFAR100-LT is used as in-distribution, our method improves the average AUROC by 1.23% and reduces the average FPR95 by 3.18% compared to the baseline method utilizing Outlier Exposure (OE). Code is available at github.

Defect identification method for capacitive-inductive dual-mode sensors with lift-off interference

The capacitive-inductive dual-mode probe provides a defect detection solution that allows simultaneous or sequential eddy current detection and capacitive imaging detection, known as capacitive-inductive dual-mode detection. The dual-mode detection combines the advantages of the two detection techniques and can be used to detect multiple types of defects in “insulator-conductor” multilayer hybrid structures. However, its high sensitivity poses a challenge as the probe's responses a susceptible to interference from lift-off variations, leading to non-relevant indications in the results that is not indicative of defects or anomalies. In this study, the Lift-off Characteristics (LoC) of the hybrid structures are studied, and an effective method for identifying relevant indications such as defect and anomaly for dual-mode sensor response signals with lift-off interference by combining the testing data of both detection modes and the LoC curve is proposed. A series of validation experiments were conducted to demonstrate the effectiveness of the proposed method in defect identification and classification. The results confirmed that the method effectively identified defect from lift-off interference and significantly reduced the occurrence of false judgments, thereby enhancing the reliability of the dual mode detection. Furthermore, the feasibility of utilizing the method to identify relevant indication when the sensor is tilted was also demonstrated.

Evaluation of Enhanced Resnet-50 Based Deep Learning Classifier for Tomato Leaf Disease Detection and Classification

This research presents a comprehensive assessment of a sophisticated approach for the accurate detection and classification of various tomato leaf diseases using an improved ResNet-50 based deep learning classifier. The alarming increase in plant diseases has prompted the need for advanced technologies that can promptly identify and categorize these ailments to ensure agricultural sustainability. The proposed method harnesses the potential of deep convolutional neural networks (CNNs) and builds upon the ResNet-50 architecture, renowned for its depth and performance. However, the approach's innovation lies in the incorporation of enhancements such as advanced data augmentation techniques and transfer learning from a vast plant disease dataset. These modifications empower the model to learn intricate disease-specific features and patterns, leading to heightened accuracy and robustness. The evaluation of the approach is conducted on an extensive dataset encompassing high-resolution images of tomato leaves affected by a range of diseases. The dataset is meticulously preprocessed to ensure consistency and quality, followed by a rigorous training regimen that fine-tunes the improved ResNet-50 model.The results underscore the efficacy of the proposed method in accurate disease detection and classification. The improved ResNet-50 based classifier demonstrates exceptional performance, achieving an impressive accuracy exceeding 95%. Notably, the model showcases resilience against variations in lighting conditions, angles, and disease severity, highlighting its applicability in real-world agricultural scenarios. The implications of this research are significant, offering an efficient and reliable tool for early detection and classification of tomato leaf diseases. The integration of advanced deep learning techniques and the enhancements introduced in this work signify a substantial advancement in precision agriculture and sustainable crop management practices. As future work, this approach can be extended to address diseases in other plant species, contributing to a versatile framework that can safeguard global food production and alleviate the challenges posed by plant diseases.

Quantum K‐Nearest Neighbor Classification Algorithm via a Divide‐and‐Conquer Strategy

AbstractThe K‐nearest neighbor algorithm is one of the most frequently applied supervised machine learning algorithms. Similarity computing is considered to be the most crucial and time‐consuming step among the classical K‐nearest neighbor algorithm. A quantum K‐nearest neighbor algorithm is proposed based on the divide‐and‐conquer strategy. A quantum circuit is designed to calculate the fidelity between the test sample and each feature vector of the training dataset. The quantum K‐nearest neighbor algorithm has higher classification efficiency in high‐dimensional data processing. The classification accuracy of the proposed algorithm is equivalent to that of the classical K‐nearest neighbor algorithm under the IRIS dataset. In addition, compared with the typical quantum K‐nearest neighbor algorithms, the proposed classification method possesses higher classification accuracy with less calculation time, which has wide applications in the industrial field.

A distance-based kernel for classification via Support Vector Machines.

Support Vector Machines (SVMs) are a type of supervised machine learning algorithm widely used for classification tasks. In contrast to traditional methods that split the data into separate training and testing sets, here we propose an innovative approach where subsets of the original data are randomly selected to train the model multiple times. This iterative training process aims to identify a representative data subset, leading to improved inferences about the population. Additionally, we introduce a novel distance-based kernel specifically designed for binary-type features based on a similarity matrix that efficiently handles both binary and multi-class classification problems. Computational experiments on publicly available datasets of varying sizes demonstrate that our proposed method significantly outperforms existing approaches in terms of classification accuracy. Furthermore, the distance-based kernel achieves superior performance compared to other well-known kernels from the literature and those used in previous studies on the same datasets. These findings validate the effectiveness of our proposed classification method and distance-based kernel for SVMs. By leveraging random subset selection and a unique kernel design, we achieve notable improvements in classification accuracy. These results have significant implications for diverse classification problems in Machine Learning and data analysis.

A bitcoin service community classification method based on Random Forest and improved KNN algorithm

AbstractThere are service communities with different functions in the Bitcoin transactions system. Identifying community categories helps to further understand the Bitcoin transactions system and facilitates targeted regulation of anonymized Bitcoin transactions. To this end, a Bitcoin service community classification method based on Random Forest and improved K‐Nearest Neighbor (KNN) algorithm is proposed. First, the transaction characteristics of different types of communities are analyzed and summarized, and the corresponding transaction features are extracted from the address and entity levels; then multiple classification algorithms are compared, the optimal model to filter the effective features is selected, and the feature vector of entity addresses is constructed. Finally, a classification model is constructed based on Random Forest and improved KNN algorithm to classify the entities. By constructing different classification models for experimental comparison, the accuracy and stability advantages of the proposed method for classification in service community classification research are verified.

Thermal kinetic parameters based classification method of residual waste oriented for efficient energy utilization

Thermal kinetic heterogeneity of mixed residual waste (RW) is a serious problem leading to unsatisfactory efficiency and stability during energy utilization processes. This work proposed a thermal kinetic parameters based classification method of RW. Considering the significant relevance between kinetic parameters and elemental compositions, the latter were used as indicators to establish the sorting principle. The proximate analysis results and kinetic parameters among the new categories were analyzed, the kinetic heterogeneity reduction performance versus the traditional classification method was compared, and the working mechanism of the proposed classification method was discussed. Three new RW categories were obtained, and the composition and kinetic parameters among these new categories have significant difference (P < 0.05). The kinetic heterogeneity was reduced compared with the traditional classification method. The absolute value of Pearson correlation coefficient between elemental composition and kinetic parameters could reach 0.8, proving the mechanism of the new classification method. The study is hopeful to facilitate efficient energy utilization of RW.

PENERAPAN METODE NAÏVE BAYES DALAM KLASIFIKASI KESEGARAN IKAN BERDASARKAN WARNA PADA CITRA AREA MATA

As a maritime nation, fish is a staple in the Indonesian diet, rich in nutrition and a crucial protein source. It is imperative to maintain the freshness of fish to ensure the quality of fish production. However, the practice of mixing fresh and non-fresh fish poses a serious threat to consumer health and diminishes the overall quality of fish production. Therefore, the development of an automated and efficient method is necessary to distinguish between fresh and non-fresh fish. This research proposes the application of the Naïve Bayes method in classifying fish freshness based on color analysis in the eye area image. This approach involves the extraction of entropy features after segmenting fish images using the RGB and YCbCr color models. A total of 40 datasets of fish eye images were used for training and testing the model. The research results indicate that the proposed classification method achieved an accuracy rate of 97.5%. This success signifies the potential of the color analysis method and entropy features in distinguishing levels of fish freshness. These findings contribute to the development of automated techniques for monitoring and processing fish quality in the fisheries industry.

A Trustworthy Curriculum Learning Guided Multi-Target Domain Adaptation Network for Autism Spectrum Disorder Classification.

Domain adaptation has demonstrated success in classification of multi-center autism spectrum disorder (ASD). However, current domain adaptation methods primarily focus on classifying data in a single target domain with the assistance of one or multiple source domains, lacking the capability to address the clinical scenario of identifying ASD in multiple target domains. In response to this limitation, we propose a Trustworthy Curriculum Learning Guided Multi-Target Domain Adaptation (TCL-MTDA) network for identifying ASD in multiple target domains. To effectively handle varying degrees of data shift in multiple target domains, we propose a trustworthy curriculum learning procedure based on the Dempster-Shafer (D-S) Theory of Evidence. Additionally, a domain-contrastive adaptation method is integrated into the TCL-MTDA process to align data distributions between source and target domains, facilitating the learning of domain-invariant features. The proposed TCL-MTDA method is evaluated on 437 subjects (including 220 ASD patients and 217 NCs) from the Autism Brain Imaging Data Exchange (ABIDE). Experimental results validate the effectiveness of our proposed method in multi-target ASD classification, achieving an average accuracy of 71.46% (95% CI: 68.85% - 74.06%) across four target domains, significantly outperforming most baseline methods (p<0.05).

Crop Classification in Mountainous Areas Using Object-Oriented Methods and Multi-Source Data: A Case Study of Xishui County, China

Accurate crop mapping can represent the fundamental data for digital agriculture and ecological security. However, current crop classification methods perform poorly in mountainous areas with small cropland field parcel areas and multiple crops under cultivation. This study proposed a new object-oriented classification method to address this issue, using multi-source data and object features to achieve multi-crop classification in mountainous areas. Firstly, a deep learning method was employed to extract cropland field parcels in mountainous areas. Subsequently, the fusion of multi-source data was carried out based on cropland field parcels, while object features tailored for mountainous crops were designed for crop classification. Comparative analysis indicates that the proposed classification method demonstrates exceptional performance, enabling accurate mapping of various crops in mountainous regions. The F1 score and overall accuracy (OA) of the proposed method are 0.8449 and 0.8502, representing a 10% improvement over the pixel-based random forest classification results. Furthermore, qualitative analysis reveals that the proposed method exhibits higher classification accuracy for smaller plots and more precise delineation of crop boundaries. Finally, meticulous crop mapping of corn, sorghum, rice, and other crops in Xishui County, Guizhou Province, demonstrates the significant potential of the proposed method in crop classification within mountainous scenarios.

Diagnosis of photovoltaic faults using digital twin and PSO-optimized shifted window transformer

This work proposes a new method for the detection, localization, and classification of grid-connected photovoltaic (PV) array faults. Line-to-line, open-module, shorted-module, open-string, and shorted-string faults as well as partial shading conditions are studied. The proposed method has two stages, which are (1) detection and localization of faults and (2) classification of faults. In the first stage, detection and localization are performed using a digital twin (DT) by analyzing the current ratio of each PV array. The measured DC (direct current) power after the operation of the DC/DC boost converter in the physical object is firstly converted into a 2-dimensional image using a recurrence plot (RP) and is then inputted to the second stage. A deep learning-based shifted windows (swin) transformer optimized by particle swarm optimization (PSO) is used in the classification stage, eliminating the need for model tuning by trial-and-error. A PV system of ten arrays with 49 kW is studied. The coefficients of determination (R2) between the results of the digital object (digital twin) and the physical object for different scenarios demonstrate the accuracy and success of the digital twin. R2 values of 0.99988 for varying irradiation with constant temperature and 0.97923 for constant irradiation with varying temperature indicate strong correlations between the digital and physical objects, further confirming the applicability of the digital twin in PV fault detection. The comparative evaluation of the PSO-optimized swin transformer against classical machine learning algorithms and state-of-the-art convolutional neural networks reveals the superior performance of the proposed method. It achieves an outstanding classification accuracy of 98.55%, demonstrating its ability to effectively classify various types of PV faults. The results of the area under the curve (AUC) of the receiver operating characteristic (ROC) curve, which measures the trade-off between true positive rate (TPR) and false positive rate (FPR), further illustrate the effectiveness of the proposed method for PV fault classification.

Traffic Classification Model in Software-Defined Networks with Artificial Intelligence Elements

Application classification is essential to improve network performance. However, with the constant growth in the number of users and applications, as well as the scaling of networks, traditional classification methods cannot fully cope with the identification and classification of network applications with the required level of delay. The use of deep learning technology together with the architecture features of software-defined networks (SDN) will allow the implementation of a new hybrid deep neural network for application classification, which can provide high classification accuracy without manual selection and feature extraction. The proposed structure proposes a classification of applications, taking into account the logical centralized management on the SDN controller. The processed data is used to train a hybrid deep neural network consisting of stacked autoencoder with a high dimensionality of the hidden layer and an output layer based on softmax regression. The necessary network flow parameters can be obtained by processing traffic with a stacked auto-encoder instead of manual processing. The softmax regression layer is used as the final application classifier. The article presents simulation results that demonstrate the advantages of the proposed classification method in comparison with the support vector machine.

Adaptive SV-Borderline SMOTE-SVM algorithm for imbalanced data classification

In recent years, imbalanced data classification has emerged as a challenging task. To address this issue, we propose an adaptive SV-Borderline SMOTE-SVM (Synthetic Minority Oversampling Technique-Support Vector Machine) algorithm, specifically designed to overcome the challenges associated with imbalanced data classification. The algorithm begins by mapping the dataset into the kernel space using SVM to identify the class boundary samples, known as support vectors (SVs). Subsequently, the neighbors of positive sample’s support vector (SV+) are calculated based on the kernel distance. Based on the class distribution of these neighbors, the SV+ samples are labeled as either “concave” or “convex”. Based on these labels, new samples are adaptively generated using two distinct calculation approaches for different labeled SV+ samples. To construct the SVM decision function without requiring the explicit expression of new samples in the kernel space, a Gram matrix is designed. Notably, all the processes ensure the credibility and reliability of the new samples. Additionally, the adaptive interpolation approach helps to ensure the security and diversity of new samples. Extensive experiments were conducted on a set of 50 KEEL datasets to evaluate the performance of our proposed method for imbalanced data classification. In experiments, our method achieved the highest G-mean score in 33 datasets and the highest F-values in 32 datasets. These results highlight the effectiveness and superiority of our proposed method compared to other approaches in addressing the challenges of imbalanced data classification.

Motor Imagery EEG Classification Based on a Weighted Multi-branch Structure Suitable for Multisubject Data.

Electroencephalogram (EEG) signal recognition based on deep learning technology requires the support of sufficient data. However, training data scarcity usually occurs in subject-specific motor imagery tasks unless multisubject data can be used to enlarge training data. Unfortunately, because of the large discrepancies between data distributions from different subjects, model performance could only be improved marginally or even worsened by simply training on multisubject data. This paper proposes a novel weighted multi-branch (WMB) structure for handling multisubject data to solve the problem, in which each branch is responsible for fitting a pair of source-target subject data and adaptive weights are used to integrate all branches or select branches with the largest weights to make the final decision. The proposed WMB structure was applied to six well-known deep learning models (EEGNet, Shallow ConvNet, Deep ConvNet, ResNet, MSFBCNN, and EEG_TCNet) and comprehensive experiments were conducted on EEG datasets BCICIV-2a, BCICIV-2b, high gamma dataset (HGD) and two supplementary datasets. Superior results against the state-of-the-art models have demonstrated the efficacy of the proposed method in subject-specific motor imagery EEG classification. For example, the proposed WMB_EEGNet achieved classification accuracies of 84.14%, 90.23%, and 97.81% on BCICIV-2a, BCICIV-2b and HGD, respectively. It is clear that the proposed WMB structure is capable to make good use of multisubject data with large distribution discrepancies for subject-specific EEG classification.

Reducing carbon footprint and promoting resource sustainability in the retail industry through the prevention of dead stocks

In the retail industry, majority of the traditional inventory management methods commonly focus on minimizing business costs but fail to consider implications on environmental sustainability. Existing inventory classification methods have failed to consider the interaction among the different performance indicators for a timely recognition of the potential dead stocks and dead stocks in the inventory. This study aims to identify and effectively classify potential dead stocks and dead stocks through an enhanced inventory classification method that integrates the remaining product shelf life as an indicator for timely recognition of dead stocks. The study also aims to provide guidelines in setting up decision rules needed in objectively classifying inventory based on each performance indicator through an automated algorithm with decision rules. Using data sets taken from open databases that provide real-world data, results show that the automated algorithm can effectively detect the presence of dead stocks and distinguish such inventory from potential dead stocks in the system which existing inventory classification methods are unable to do. One-Way ANOVA Tests performed showed that the proposed classification method can consistently classify inventory across different retail data sets while Monte Carlo simulation was used to simulate the amount of inventory that becomes dead stocks in a multi-period inventory management system. Results show that 17% of these potential dead stocks, if undetected, become dead stocks in the following period. Over a ton of carbon emissions is prevented with lesser dead stock waste disposal. The value of this study lies in the fact that classifying inventory to detect the presence of potential dead stocks is an essential first step in solving the dead stock problem not just for business profitability but also for sustainable resource consumption and a sustainable environment.

Classification of building complex for the large-scale construction of distributed photovoltaics in urban buildings

The large-scale development of urban building photovoltaics (PV) has become an important avenue for cities to achieve building energy conservation, emission reduction, and carbon neutrality. Assessing the solar energy potential of urban buildings plays a crucial role in the installation and overall planning of urban building PV systems. However, due to the vast number and diverse types of urban buildings, individual building calculations can be both labor-intensive and inefficient. Therefore, batch processing of buildings after classification is a feasible approach. Previous assessments of solar energy potential in building clusters mainly focused on typical clusters or standard geometric areas, but all urban building types were covered. Based on this, this paper proposes a method for defining Building Complexes with a Single Land Use Nature (BCSLUN) by combining geographical distribution and urban land use characteristics. Furthermore, population density, building density, compactness, and land use characteristics are selected as features, and a method that combines K-means clustering and random forest classification is proposed to complete the classification of building complexes. Taking the central urban area of Xi'an, China as an example, 5,804 building complexes are classified into four building types. The cross-validation classification accuracy is 0.99, and the Kappa coefficient is 0.94. Based on Rhino modeling, the global solar radiation on the surface of each building complex was calculated. The annual surface global radiation of the four building types is 11,442.73 MWh, 2,734.71 MWh, 13,923.81 MWh, and 67,215 MWh. The violin-box plot shows a significant classification effect, which proves the rationality of the proposed classification method. Finally, a potential map based on solar radiation from urban building surfaces is generated. Building complex classification can quickly and accurately estimate the solar energy potential of urban buildings, providing decision-making support for the large-scale construction of urban building PV. Moreover, this method can be easily applied to other cities.

Classification of inter-turn short-circuit faults in induction motors based on quaternion analysis

Short-circuit in three-phase engines are highly destructive faults, which overheat and damage internal elements reducing efficiency and lifetime. New multi-class approaches are best trained with measurements from three-phase motors instrumented with short-circuit faults because it offers natural and physical signal behavior. This work overcomes the lack of datasets by acquiring current signals from an instrumented induction motor to create a dataset of inter-turn short-circuit (ITSC) faults at four levels per phase. The dataset generated consists of 13 categories with five repetitions per trial for a squirrel cage motor induction. The proposed classification method is based on quaternions that simultaneously model the three-phase signals as pure quaternions. Three statistical features are extracted from quaternions, and a decision tree classifier is trained per feature. Thereby, a boosting scheme is used to calculate the resulting category. Boosting method improves the classification results of decision tree models, showing fast, accurate, and robust performance.

A text classification-based approach for evaluating and enhancing the machine interpretability of building codes

Interpreting regulatory documents or building codes into computer-processable formats is essential for the intelligent design and construction of buildings and infrastructures. Although automated rule interpretation (ARI) methods have been investigated for years, most of them are highly dependent on the early and manual filtering of interpretable clauses from a building code. While few of them considered machine interpretability, which represents the potential to be transformed into a computer-processable format, from both clause- and document-level. Therefore, this research aims to propose a novel approach to automatically evaluate and enhance the machine interpretability of single clauses and building codes. First, a few categories are introduced to classify each clause in a building code considering the requirements for rule interpretation, and a dataset is developed for model training. Then, an efficient text classification model is developed based on a pretrained domain-specific language model and transfer learning techniques. Finally, a quantitative evaluation method is proposed to assess the overall interpretability of building codes. Experiments show that the proposed text classification algorithm outperforms the existing CNN- or RNN-based methods, by improving the F1-score from 72.16% to 93.60%. It is also illustrated that the proposed classification method can enhance downstream ARI methods with an improvement of 4%. Furthermore, analysis of more than 150 building codes in China showed that their average interpretability is only 34.40%, which implies that it is still difficult to fully transform an entire regulatory document into computer-processable formats. It is also argued that the interpretability of building codes should be further improved both from the human side (considering certain constraints when writing building codes) and the machine side (developing more powerful algorithms, tools, etc.).

A novel total phosphorus concentration retrieval method based on two-line classification in lakes and reservoirs across China

Phosphorus is widely recognized as a nutrient that restricts growth and is the primary contributor to eutrophication in 80 % of water bodies. Consequently, the Chinese government has consistently prioritized monitoring and controlling total phosphorus (TP) levels. The remote estimation of TP in lakes and reservoirs at a national scale is a challenging task due to TP being a non-optically active parameter. Currently, there is a lack of developed TP inversion models specifically designed for lakes and reservoirs in China. For solving this problem, a novel two-line classification method drawn on scatter plots based on the natural logarithm of TP (Ln(TP)) and B33/B9 was proposed and used to classify 1211 measured samples obtained from field cruises in 105 lakes and reservoirs across China from 2012 to 2022 into three categories, Class 1, Class 2, and Class 3. Results demonstrate that the proposed classification method has the ability to enhance the correlation between Ln(TP) and 43 basic potential single band and band combinations. Specifically, the correlation range improved from (−0.31,0.15) to (−0.77,0.24) in Class 1, (−0.81, 0.36) in Class 2, and (−0.74, 0.52) in Class 3. Additionally, the classification method also improved the correlation range between Ln(TP) and 820 band ratios, from (−0.32, 0.32) to (−0.83, 0.82) in Class 1, (−0.86, 0.86) in Class 2, and (−0.86, 0.86) in Class 3. These datasets were subsequently utilized as input for eXtreme Gradient Boosting (XGBoost) models. Finally, well performing XGBoost models in Class 1 (R2 = 0.76, RMSE = 0.3, MAPE = 12 %), Class 2 (R2 = 0.84, RMSE = 0.49, MAPE = 38 %), and Class 3 (R2 = 0.74, RMSE = 0.46, MAPE = 14 %) were used to map TP of 563 large lakes and reservoirs (≥20 km2) across China using MODIS images from 2005, 2010, 2015, and 2020. This study presents a novel approach for estimating non-optically active parameters through remote sensing on a national scale.