Optimal Model Configuration Research Articles

Automated Agricultural Crop Type Mapping Using Fusion of Transfer Learning and Tasmanian devil Optimization Algorithm on Remote Sensing Imagery

At present, the application of remote sensing (RS) data achieved from satellite imagery or unmanned aerial vehicles (UAV) has become common for crop classification procedures, i.e. crop mapping, soil classification, or prediction of yield. The classification of food crop utilizing RS images (RSI) is one of the major applications of RS in farming. It contains the usage of aerial or satellite images for classifying and identifying dissimilar kinds of food crops developed in an exact region. This data is beneficial for estimation of yield, crop monitoring, and land management. Meeting the conditions for examining these data needs more refined techniques and artificial intelligence (AI) technologies, which deliver essential support. Recently, the usage of deep learning (DL) for crop type classification with RS images could help sustainable farming practices by providing appropriate and precise data on the kinds and features of crops. In this study, we offer an Automated Agricultural Crop Type Mapping Utilizing Fusion of Transfer Learning and Tasmanian Devil Optimization (AACTM-FTLTDO) algorithm on Remote Sensing Imagery. The primary goal of the AACTM-FTLTDO methodology is to accurately detect and classify crop types for more precise agricultural monitoring using remote sensing technologies. To accomplish that, the AACTM-FTLTDO model employs a fusion of transfer learning techniques involving three models such as SqueezeNet, CapsNet, and ShuffleNetV2 to capture diverse, multi-scale spatial and spectral features. For the crop type classification and detection process, the auto-encoder (AE) classifier can be employed. Eventually, the tasmanian devil optimization (TDO) technique was deployed to modify the hyperparameter of the AE technique for ensuring optimal model configurations and reducing computational complexity. A wide range of experimentation studies is made and the results are examined under numerous measures. The comparative study shows that the AACTM-FTLTDO technique performs better than existing approaches

A new prediction model based on deep learning for pig house environment

A prediction model of the pig house environment based on Bayesian optimization (BO), squeeze and excitation block (SE), convolutional neural network (CNN) and gated recurrent unit (GRU) is proposed to improve the prediction accuracy and animal welfare and take control measures in advance. To ensure the optimal model configuration, the model uses a BO algorithm to fine-tune hyper-parameters, such as the number of GRUs, initial learning rate and L2 normal form regularization factor. The environmental data are fed into the SE-CNN block, which extracts the local features of the data through convolutional operations. The SE block further learns the weights of the feature channels, highlights the important features and suppresses the unimportant ones, improving the feature discrimination ability. The extracted local features are fed into the GRU network to capture the long-term dependency in the sequence, and this information is used to predict future values. The indoor environmental parameters of the pig house are predicted. The prediction performance is evaluated through comparative experiments. The model outperforms other models (e.g., CNN-LSTM, CNN-BiLSTM and CNN-GRU) in predicting temperature, humidity, CO2 and NH3 concentrations. It has higher coefficient of determination (R2), lower mean absolute error (MSE), and mean absolute percentage error (MAPE), especially in the prediction of ammonia, which reaches R2 of 0. 9883, MSE of 0.03243, and MAPE of 0.01536. These data demonstrate the significant advantages of the BO-SE-CNN-GRU model in prediction accuracy and stability. This model provides decision support for environmental control of pig houses.

Deep Learning-Based Risk Analysis and Prediction During the Implementation of Carbon Neutrality Goals

Risk prediction has become increasingly crucial in today's complex and dynamic environments. However, existing forecasting methods still face challenges in terms of accuracy and reliability. Therefore, it is imperative to explore new approaches to better address risks. In response to this need, our study introduces an innovative risk prediction model known as WOA-FPALSTM. What sets this model apart is its seamless integration of deep learning and heuristic algorithms, designed to overcome the limitations of existing approaches. The core component of deep learning, LSTM, excels in sequence modeling by capturing long-term and short-term dependencies in time series data, thereby enhancing the model's ability to model temporal data. Meanwhile, the heuristic algorithm, WOA (Whale Optimization Algorithm), equips our model with global search capabilities, facilitating the discovery of optimal model configurations and significantly improving predictive performance.

Regional evaluation of simulated waves during tropical storm events in the Gulf of Mexico

High waves in the Gulf of Mexico are mainly originated in tropical cyclones and winter storms. The availability of accurate estimations of high wave characteristics is an essential prerequisite for marine spatial planning, coastal zone management, and marine developments over the Northeastern Gulf of Mexico (NeGoM). Moreover, accurate wave simulation relies upon accurate wind field estimations. The wind field data used as input for numerical wave models is typically obtained from global and regional atmospheric model outputs. However, these model outputs may exhibit higher bias relative to observations, especially during fast-traveling tropical weather events. This study evaluates wave sensitivity over the NeGoM to different input wind speed and direction sources using unstructured SWAN model simulations. Five wind data products were employed in this study from ECMWF and NCEP global and regional models to examine the impact of wind forcing on wave simulations. The quality of wind field input data from these sources was evaluated relative to available observations (wind speed and direction) at National Data Buoy Center stations. A detailed sensitivity analysis is carried out to determine optimal wave model configuration. The model's performance is evaluated by comparing its wave simulation results with observational data during selected periods, simultaneously assessing the quality and sufficiency of wind data inputs from various sources. Our study shows that wind fields from FNL and NAM datasets are better correlated with observational data, in terms of vector correlation. Simulations using ECMWF – Real-Time product, combined with Janssen white-capping scheme better estimate wave conditions over the NeGoM compared with the other wind forcings used in this study. Moreover, Janssen white-capping scheme seems to be more realistic for estimations during Hurricane Michael; while for Hurricane Ida, Janssen generation scheme tends to overestimate the wave heights and Komen provides more accurate results.

Utilizing MicroGenetic Algorithm for Optimal Design of Permanent-Magnet-Assisted WFSM for Traction Machines

With increasing worries about the environment, there is a rising focus on saving energy in various industries. In the e-mobility industry of electric motors, permanent magnet synchronous motors (PMSMs) are widely utilized for saving energy due to their high-efficiency motor technologies. However, challenges like environmental degradation from rare earth development and difficulties in controlling magnetic field fluctuations persist. To address these issues, active research focuses on the wound field synchronous motor (WFSM), known for its ability to regulate field current efficiently across various speeds and operating conditions. Nevertheless, compared with other synchronous motors, the WFSM tends to exhibit relatively lower efficiency and torque density. Because the WFSM involves winding both the rotor and the stator, it results in increased copper and iron losses. In this article, a model that enhances torque density by inserting permanent magnets (PMs) into the rotor of the basic WFSM is proposed. This proposed model bolsters the d axis magnetic flux, thereby enhancing the motor’s overall performance while addressing environmental concerns related to rare-earth materials and potentially reducing manufacturing costs when compared with those of the PMSM. The research methodology involves a comprehensive sensitivity analysis to identify key design variables, followed by sampling using optimal Latin hypercube design (OLHD). A surrogate model is then constructed using the kriging interpolation technique, and the optimization process employs a micro-genetic algorithm (MGA) to derive the optimal model configuration. The algorithm was performed to minimize the use of PMs when the same torque as that of the basic WFSM is present, and to reduce torque ripple. Error assessment is conducted through comparisons with finite element method (FEM) simulations. The optimized permanent-magnet-assisted WFSM (PMa-WFSM) model improved efficiency by 1.08% when it was the same size as the basic WFSM, and the torque ripple decreased by 5.43%. The proposed PMa-WFSM derived from this article is expected to be suitable for use in the e-mobility industry as a replacement for PMSM.

Understanding the Role of Self-Attention in a Transformer Model for the Discrimination of SCD From MCI Using Resting-State EEG.

The identification of EEG biomarkers to discriminate Subjective Cognitive Decline (SCD) from Mild Cognitive Impairment (MCI) conditions is a complex task which requires great clinical effort and expertise. We exploit the self-attention component of the Transformer architecture to obtain physiological explanations of the model's decisions in the discrimination of 56 SCD and 45 MCI patients using resting-state EEG. Specifically, an interpretability workflow leveraging attention scores and time-frequency analysis of EEG epochs through Continuous Wavelet Transform is proposed. In the classification framework, models are trained and validated with 5-fold cross-validation and evaluated on a test set obtained by selecting 20% of the total subjects. Ablation studies and hyperparameter tuning tests are conducted to identify the optimal model configuration. Results show that the best performing model, which achieves acceptable results both on epochs' and patients' classification, is capable of finding specific EEG patterns that highlight changes in the brain activity between the two conditions. We demonstrate the potential of attention weights as tools to guide experts in understanding which disease-relevant EEG features could be discriminative of SCD and MCI.

Design of Software Reliability Prediction using Radial Basis Function Networks with Nonlinear Analysis and Topological Considerations

This study presents a thorough methodology for predicting software reliability by employing Radial Basis Function Networks (RBFNs). By using machine learning methodologies such as Radial Basis Function Networks (RBFNs), software development teams are enabled to make well-informed decisions, optimize resource allocation, and proactively mitigate potential dependability concerns. Consequently, this results in improved software quality and heightened customer happiness. This methodology involves multiple stages, starting with data collection and preprocessing. We assemble a comprehensive dataset comprising historical software performance metrics, defect reports, and relevant development process information. After data cleansing and feature engineering, we split the dataset into training, validation, and testing subsets. The core of this approach lies in the construction and training of RBFNs. These neural networks consist of an input layer, a hidden layer with radial basis functions, and an output layer. The architecture parameters, particularly the number of hidden neurons and the spread parameter of the radial basis functions, are optimized through a systematic hyperparameter tuning process using the validation dataset. Upon achieving an optimal model configuration, we rigorously evaluate the RBFN predictive capabilities using the testing dataset.

Machine learning in epidemiology: Neural networks forecasting of monkeypox cases.

This study integrates advanced machine learning techniques, namely Artificial Neural Networks, Long Short-Term Memory, and Gated Recurrent Unit models, to forecast monkeypox outbreaks in Canada, Spain, the USA, and Portugal. The research focuses on the effectiveness of these models in predicting the spread and severity of cases using data from June 3 to December 31, 2022, and evaluates them against test data from January 1 to February 7, 2023. The study highlights the potential of neural networks in epidemiology, especially concerning recent monkeypox outbreaks. It provides a comparative analysis of the models, emphasizing their capabilities in public health strategies. The research identifies optimal model configurations and underscores the efficiency of the Levenberg-Marquardt algorithm in training. The findings suggest that ANN models, particularly those with optimized Root Mean Squared Error, Mean Absolute Percentage Error, and the Coefficient of Determination values, are effective in infectious disease forecasting and can significantly enhance public health responses.

Assessment on fast simulation of wind-driven pollutant dispersion around a street canyon with regime-switching Markov chain

High-density cities shorten commutes but also create proximity to air pollution sources. Sensors coupled with simulations can achieve real-time monitoring and risk assessment, and the recently emerging data-driven models provide opportunities that allow minimizing the computation burden of physical models. This study investigates the use of a time inhomogeneous Markov chain (CLF-MC) to fast simulation of wind-driven pollutant dispersion around a street canyon. The CLF-MC model is trained and tested using synthetic datasets obtained from large eddy simulations (LES). The LES model is validated using a benchmark wind tunnel test, which yields correlation coefficients above 0.95. The results show that the CLF-MC can faithfully reproduce the cumulative distribution of air pollution concentrations at the receptor site using the optimal model configuration identified in this study. For the hold-out test dataset, the release condition is a time-varying release strength, and the CLF-MC shows improved prediction results compared to the baseline models, i.e. the weighted average deviation is 15 % and 51 % smaller, respectively. In comparison to LES, the CLF-MC model is 4.8 × 106 times faster. These findings provide scientific references for fast and accurate wind-driven pollutant dispersion simulation using the probabilistic modelling approach.

A Comparative Study of Vehicle Velocity Prediction for Hybrid Electric Vehicles Based on a Neural Network

Vehicle velocity prediction (VVP) plays a pivotal role in determining the power demand of hybrid electric vehicles, which is crucial for establishing effective energy management strategies and, subsequently, improving the fuel economy. Neural networks (NNs) have emerged as a powerful tool for VVP, due to their robustness and non-linear mapping capabilities. This paper describes a comprehensive exploration of NN-based VVP methods employing both qualitative theory analysis and quantitative numerical simulations. The used methodology involved the extraction of key feature parameters for model inputs through the utilization of Pearson correlation coefficients and the random forest (RF) method. Subsequently, three distinct NN-based VVP models were constructed comprising the following: a backpropagation neural network (BPNN) model, a long short-term memory (LSTM) model, and a generative pre-training (GPT) model. Simulation experiments were conducted to investigate various factors, such as the feature parameters, sliding window length, and prediction horizon, and the prediction accuracy and computation time were identified as key performance metrics for VVP. Finally, the relationship between the model inputs and velocity prediction performance was revealed through various comparative analyses. This study not only facilitated the identification of an optimal NN model configuration to balance prediction accuracy and computation time, but also serves as a foundational step toward enhancing the energy efficiency of hybrid electric vehicles.

A Model for Predicting Short‐Term Operating Speeds of Compact Passenger Vehicles on Interchange Ramps Within Urban Expressway Networks

The prediction of operating speed plays a crucial role in road design and safety assessment, especially on complex urban expressway interchange ramps. This task is challenging due to various influences like road conditions, traffic dynamics, and driver behavior. This study aims to identify the optimal model configuration for predicting operating speeds on urban expressway interchange ramps. Three models are established: a short‐term operating speed model based on a generalized linear model (GLM), a GLM incorporating for spatial correlation (GLMS), and a deep neural network model considering spatial correlation (DNNS). Each model incorporates considerations for the impact of the plan, profile, and other facets of the interchange ramp in urban expressways. Naturalistic driving experiments are conducted in Shanghai, 70% for model calibration and 30% for validation. Comparative analysis shows that the DNNS model outperforms the others, effectively capturing speed fluctuations along the interchange ramp, demonstrating its robustness and generalization capabilities.

Feature Engineering and Model Optimization Based Classification Method for Network Intrusion Detection

In light of the escalating ubiquity of the Internet, the proliferation of cyber-attacks, coupled with their intricate and surreptitious nature, has significantly imperiled network security. Traditional machine learning methodologies inherently exhibit constraints in effectively detecting and classifying multifarious cyber threats. Specifically, the surge in high-dimensional network traffic data and the imbalanced distribution of classes exacerbate the predicament of ideal classification performance. Notably, the presence of redundant information within network traffic data undermines the accuracy of classifiers. To address these challenges, this study introduces a novel approach for intrusion detection classification which integrates advanced techniques of feature engineering and model optimization. The method employs a feature engineering approach that leverages mutual information maximum correlation minimum redundancy (mRMR) feature selection and synthetic minority class oversampling technique (SMOTE) to process network data. This transformation of raw data into more meaningful features effectively addresses the complexity and diversity inherent in network data, enhancing classifier accuracy by reducing feature redundancy and mitigating issues related to class imbalance and the detection of rare attacks. Furthermore, to optimize classifier performance, the paper applies the Optuna method to fine-tune the hyperparameters of the Catboost classifier, thereby determining the optimal model configuration. The study conducts binary and multi-classification experiments using publicly available datasets, including NSL_KDD, UNSW-NB15, and CICIDS-2017. Experimental results demonstrate that the proposed method outperforms traditional approaches regarding accuracy, recall, precision, and F-value. These findings highlight the method’s potential and performance in network intrusion detection.

Automatic detection of prostate cancer grades and chronic prostatitis in biparametric MRI

Background and objective:With emerging evidence to improve prostate cancer (PCa) screening, multiparametric magnetic prostate imaging is becoming an essential noninvasive component of the diagnostic routine. Computer-aided diagnostic (CAD) tools powered by deep learning can help radiologists interpret multiple volumetric images. In this work, our objective was to examine promising methods recently proposed in the multigrade prostate cancer detection task and to suggest practical considerations regarding model training in this context.Methods:We collected 1647 fine-grained biopsy-confirmed findings, including Gleason scores and prostatitis, to form a training dataset. In our experimental framework for lesion detection, all models utilized 3D nnU-Net architecture that accounts for anisotropy in the MRI data. First, we explore an optimal range of b-values for diffusion-weighted imaging (DWI) modality and its effect on the detection of clinically significant prostate cancer (csPCa) and prostatitis using deep learning, as the optimal range is not yet clearly defined in this domain. Next, we propose a simulated multimodal shift as a data augmentation technique to compensate for the multimodal shift present in the data. Third, we study the effect of incorporating the prostatitis class alongside cancer-related findings at three different granularities of the prostate cancer class (coarse, medium, and fine) and its impact on the detection rate of the target csPCa. Furthermore, ordinal and one-hot encoded (OHE) output formulations were tested.Results: An optimal model configuration with fine class granularity (prostatitis included) and OHE has scored the lesion-wise partial Free-Response Receiver Operating Characteristic (FROC) area under the curve (AUC) of 1.94 (CI 95%: 1.76–2.11) and patient-wise ROC AUC of 0.874 (CI 95%: 0.793–0.938) in the detection of csPCa. Inclusion of the auxiliary prostatitis class has demonstrated a stable relative improvement in specificity at a false positive rate (FPR) of 1.0 per patient, with an increase of 3%, 7%, and 4% for coarse, medium, and fine class granularities.Conclusions: This paper examines several configurations for model training in the biparametric MRI setup and proposes optimal value ranges. It also shows that the fine-grained class configuration, including prostatitis, is beneficial for detecting csPCa. The ability to detect prostatitis in all low-risk cancer lesions suggests the potential to improve the quality of the early diagnosis of prostate diseases. It also implies an improved interpretability of the results by the radiologist.

Study on Driver Cross-Subject Emotion Recognition Based on Raw Multi-Channels EEG Data

In our life, emotions often have a profound impact on human behavior, especially for drivers, as negative emotions can increase the risk of traffic accidents. As such, it is imperative to accurately discern the emotional states of drivers in order to preemptively address and mitigate any negative emotions that may otherwise manifest and compromise driving behavior. In contrast to many current studies that rely on complex and deep neural network models to achieve high accuracy, this research aims to explore the potential of achieving high recognition accuracy using shallow neural networks through restructuring the structure and dimensions of the data. In this study, we propose an end-to-end convolutional neural network (CNN) model called simply ameliorated CNN (SACNN) to address the issue of low accuracy in cross-subject emotion recognition. We extracted features and converted dimensions of EEG signals from the SEED dataset from the BCMI Laboratory to construct 62-dimensional data, and obtained the optimal model configuration through ablation experiments. To further improve recognition accuracy, we selected the top 10 channels with the highest accuracy by separately training the EEG data of each of the 62 channels. The results showed that the SACNN model achieved an accuracy of 88.16% based on raw cross-subject data, and an accuracy of 91.85% based on EEG channel data from the top 10 channels. In addition, we explored the impact of the position of the BN and dropout layers on the model through experiments, and found that a targeted shallow CNN model performed better than deeper and larger perceptual field CNN models. Furthermore, we discuss herein the future issues and challenges of driver emotion recognition in promising smart city applications.

Pseudo-Label Guided Sparse Deep Belief Network Learning Method for Fault Diagnosis of Radar Critical Components

Effective fault diagnosis of critical components is essential to ensure the safe and reliable operation of the entire system. This paper deals with the fault diagnosis of transmitter/receiver module, which is a critical component in the phased array radar system, by proposing a novel deep belief network learning method. A sparse deep belief network based on Gaussian function is first constructed to automatically learn the relationship between monitoring data and component health conditions. With the trained sparse deep belief network, the pseudo-labels are produced for unlabeled samples, while the information entropy is employed to calculate the confidence levels reflecting their certainty to reduce the effect of pseudo-label noise. The pseudo-labeled samples with high confidence levels are added to the training set to retrain the network. Optimal model configuration parameters are obtained through a chaos game optimization algorithm. The effectiveness of the proposed method is verified on a real-world dataset from a certain type of phased array radar. The experiments show that the mean identification rate of this method can reach 96.33%, which not only exceeds some deep belief network-based modeling methods, but also exceeds other intelligent methods.

Multi Kelas Speaker Recognition Menggunakan Deep Learning dengan CN-Celeb Dataset

Speaker recognition has been widely applied in various fields of human life such as Siri from Apple, Cortana from Microsoft, and Voice Assistant by Google. One of the problems when creating speaker recognition is related to the dataset used for the modeling process. The dataset used for creating the speaker recognition model is mostly data that cannot represent real-world conditions. The result is when implemented in the real-world conditions are not optimal. This study develops a speaker recognition model using deep learning (LSTM) with the CN-Celeb dataset. The CN-Celeb dataset is data taken directly from the real world so there is a lot of noise. The hope of using this dataset is that it can represent real world conditions. Model development uses 2 stacked LSTM for multi-class speaker recognition tasks. In addition, this study performs tuning hyperparameters with a grid search method to obtain the most optimal model configuration. The results showed that the EER value of the LSTM model was 10.13% better than the reference baseline paper of 15.52%. In addition, when compared with other studies that also used the CN-Celeb dataset but using different models, it was found that the LSTM model had promising results. From the results of study that has been carried out and also compared with other people's research, it was found that the LSTM model gave promising performance. The LSTM model is compared with the x-vectors, PLDA, TDNN, and transformers models

Sensitivity analysis of a three-dimensional simulation of turbidity currents in a sloping flume

Sensitivity analysis is necessary in numerical models to establish the optimal model configuration and simulate turbidity currents accurately. The sensitivity of turbidity currents simulated by a three-dimensional numerical model named the semi-implicit cross-scale hydroscience integrated system model (SCHISM) was assessed using domain discretization, the turbulence closure model, and the transport scheme. The results show that a higher horizontal resolution with an appropriate aspect ratio, localized sigma coordinates with shaved cells (LSC2) grid with sufficient layers, a two-phase mixture turbulence closure model, and transport schemes with critical depth ratio lower than 1 could improve the model performance on the turbidity currents. The study can provide helpful guidance to establish accurate turbidity current simulations in complex water environments.

Deep learning modeling of public’s sentiments towards temporal evolution of COVID-19 transmission

Public sentiments towards global pandemics are important for public health assessment and disease control. This study develops a modularized deep learning framework to quantify public sentiments towards COVID-19, followed by leveraging the predicted sentiments to model and forecast the daily growth rate of confirmed COVID-19 cases globally, via a proposed G parameter. In the proposed framework, public sentiments are first modeled via a valence dimensional indicator, instead of discrete schemas, and are classified into 4 primary emotional categories: (a) neutral; (b) negative; (c) positive; (d) ambivalent, by using multiple word embedding models and classifiers for text sentiments analyses and classification. The trained model is subsequently applied to analyze large volumes (millions in quantity) of daily Tweets pertaining to COVID-19, ranging from 22 Jan 2020 to 10 May 2020. The results demonstrate that the global community gradually evokes both positive and negative sentiments towards COVID-19 over time compared to the dominant neural emotion at its inception. The predicted time-series sentiments are then leveraged to train a deep neural network (DNN) to model and forecast the G parameter by achieving the lowest possible mean absolute percentage error (MAPE) score of around 17.0% during the model’s testing step with the optimal model configuration.

Development of smart aquaculture farm management system using IoT and AI-based surrogate models

Due to low labor participation by young adults and an aging agricultural population, Taiwan and the rest of the world are facing labor shortages in agriculture, which will affect aquaculture production. The proposed system is intended primarily for solving the problems faced by the aquaculture farming sector in Taiwan by designing a smart IoT-based fish monitoring and control system equipped with different IoT devices to enable real-time data collection; so that fishpond water-quality conditions and other system parameters can be readily monitored, adjusted, and assessed remotely. To predict the growth of the California Bass fish, this study also develops a deep learning model (DL) that correlates the different parameters of the smart aquaculture system. Bayesian optimization-based hyper-parameter tuning was employed to find the optimal DL model configuration to produce accurate predictions on the given experimental data set. The optimal model produces an R 2 value of 0.94 and a mean square error of 0.0015, demonstrating the applicability of the model to predict the desired output. Based on the results of the experiments, the DL model can be incorporated into the autonomous feeding system, reducing the amount of leftover feed. Thus, aquaculture based on the artificial intelligence of things (AIOT) can assist fish farmers in intelligently controlling and managing different fishpond equipment remotely and assist aquaculture operators in performing professional aquaculture, lowering the industry's entry barrier, and promoting aquaculture. • Developed an AIoT-based smart aquaculture farm management system that enables remote control and monitoring. • A DL model is developed to correlate the different parameters of a fish-feeding system and predict the growth of fish. • Bayesian optimization is utilized to find the most suitable DL model that will produce the highest R2 and mean square error.

Developing a Climate Prediction System over Southwest China Using the 8-km Weather Research and Forecasting (WRF) Model: System Design, Model Calibration, and Performance Evaluation

Abstract A high-resolution, short-term climate prediction system for summer (June–July–August) climate over Southwest China has been developed using the Weather Research and Forecasting (WRF) Model nested with a global climate prediction system (PCCSM4). The system includes 12 ensemble members generated by PCCSM4 with different initial conditions, and the finest horizontal resolution of WRF is 8 km. This study evaluates the ability of the WRF Model to predict summer climate over Southwest China, focusing on the system design, model tuning, and evaluation of baseline model performance. Sensitivity simulations are first conducted to provide the optimal model configuration, and the model performance is evaluated against available observational data using reforecast simulations for 1981–2020. When compared to PCCSM4, the WRF Model shows major improvements in predicting the spatial distribution of major variables such as 2-m temperature, 10-m wind speed, and precipitation. WRF also shows better skill in predicting interannual temperature variability and extreme temperature events, with higher anomaly correlation coefficients. However, large model biases remain in seasonal precipitation anomaly predictions. Overall, this study highlights the potential advantages of using the high-resolution WRF Model to predict summer climate conditions over Southwest China.