Backpropagation Research Articles

Evaluation of rural tourism development level using BERT-enhanced deep learning model and BP algorithm

To address the insufficient expressive capabilities of traditional methods in assessing the development level of rural tourism, this study explores the fusion application of the Bidirectional Encoder Representations from Transformers (BERT) and the Back Propagation (BP) algorithm to enhance the accuracy and comprehensiveness of rural tourism development assessment. Firstly, this study introduces the BERT deep learning model and its applications in natural language processing, alongside the role of the BP algorithm in pattern recognition and predictive analysis. Subsequently, a framework for assessing rural tourism development levels, integrating BERT and the BP algorithm, is proposed. This framework collects multidimensional rural tourism-related data and utilizes the BERT model for sentiment analysis and topic extraction from textual data. Empirical analysis of rural tourism development in a specific region validates the effectiveness of the proposed approach. Experimental results demonstrate: (1) The model achieves an accuracy of 84.33% and an F1 score of 85.33% on the publicly available Laptop dataset, with a processing time of 20 s, significantly outperforming other methods. Compared to traditional approaches, the proposed method accurately captures correlations between textual information and numerical data, thereby enhancing the credibility and accuracy of assessment results. (2) From the ablation study results, it is evident that removing any component from the model leads to performance degradation. Specifically, removing the Bidirectional Gated Recurrent Unit (BiGRU) reduces accuracy and F1 scores to 78.21% and 76.33% on the Laptop dataset, and to 85.10% and 70.45% on the Tourist_F dataset. Removing Text Convolutional Neural Network (CNN) reduces accuracy and F1 scores to 79.34% and 77.56% on the Laptop dataset, and to 86.25% and 72.11% on the Tourist_F dataset. The most significant performance decline occurs upon removing BERT, with accuracy and F1 scores decreasing to 70.14% and 66.43% on the Laptop dataset, and to 78.45% and 60.33% on the Tourist_F dataset. These results indicate that BiGcRU, TextCNN, and BERT each contribute significantly to the model’s performance. This study provides substantial support for advancing sustainable development in rural tourism, offering significant practical and innovative value.

A Model for Fat Content Detection in Walnuts Based on Near-Infrared Spectroscopy

Near-infrared spectroscopy (NIR) is an efficient and accurate method for fat content detection in walnuts. ‘Wen 185’ walnut is grown in large quantities in southern Xinjiang, and its fat content is an important indicator for evaluating the intrinsic quality. The excessive pursuit of yield efficiency, combined with the neglect of quality, agricultural product safety and other factors, has led to the production of poor-quality walnuts. Moreover, research on predicting walnut kernel fat content based on near-infrared spectroscopy technology is rarely reported. Therefore, a technical framework for walnut kernel detection based on near-infrared spectroscopy and the technical standards for ‘Wen 185’ are urgently needed. After first optimizing the initial spectrum data using five preprocessing methods, we established separate prediction models for walnut kernel fat content based on either a back propagation neural network or a support vector regression (SVR) algorithm. The results show that the correction set and validation set coefficients of determination of the walnut kernel fat content prediction model using the back propagation neural network algorithm were 0.86 and 0.89, respectively, with root mean square errors of 1.56 and 1.58, and an RPD value of 2.57; the coefficients of determination for the calibration and validation sets of SVR were 0.90 and 0.83, respectively, with root mean square errors of 1.76 and 1.70, respectively, and an RPD value of 1.70. Thus, near-infrared spectroscopy can be used to establish a foundation for realizing the rapid detection of walnut kernel fat content.

Modeling rate of penetration using hybridization of Artificial Neural Network and Artificial Fish Swarm algorithm

This study aims to develop a robust and efficient approach for predicting the Rate of Penetration (ROP) by integrating the Artificial Fish Swarm Algorithm (AFSA) with the Back Propagation (BP) algorithm in an Artificial Neural Network (ANN). The integration of AFSA with Back Propagation is the primary novelty of this work, as it seeks to achieve optimal weight and bias values for the ANN, thereby enhancing prediction accuracy. A dataset of 1944 data points from a vertical well in southern Algeria was utilized. The performance of the developed models was evaluated using correlation coefficient and root mean square error (RMSE). The results demonstrated that the AFSA-ANN model significantly outperforms conventional ANN, Support Vector Regression (SVR), Random Forest regressor (RFR), and Bourgoyne and Young’s model in terms of accuracy. Additionally, the application of leverage method revealed that only 0.98% of the data falls outside the model’s applicability domain, indicating that the data and the model are statistically valid and reliable. The findings of this study have significant implications for the oil and gas field, providing a more accurate and reliable method for predicting ROP, which can lead to more efficient drilling operations and reduced costs.

Mathematical simulation of tangent hyperbolic nanofluid flow coupled with homogenous/heterogeneous chemical reaction by using Levenberg–Marquardt back propagation over a Riga plate

Mathematical simulation of tangent hyperbolic nanofluid flow coupled with homogenous/heterogeneous chemical reaction by using Levenberg–Marquardt back propagation over a Riga plate

Visualization Analysis of Integrating College Sports Training and Psychology into Basketball Physical Education Teaching System Based on Image Recognition Algorithm

ABSTRACT On this basis, a basketball movement model based on ORB (Oriented FAST and Rotated BRIEF) local feature extraction has been proposed, and a basketball teaching visualization analysis system has been constructed. By extracting images from athletes’ videos in basketball, and using their training as a standard, tracking the frequency of their various celebration postures appearing and improving their scores in the game, corresponding feature points are extracted, and then the BP (back propagation) neural network algorithm is used to classify feature points. Ten different participants were selected from 10 games. Their movements were tracked, and then the extracted feature actions were imported into the visualization system for visual analysis, comparing the changes before and after psychological intervention. The results showed that there were significant differences in scores, errors, hit rates, steals, and motivation among subjects who received psychological cues and incentives compared to the data before psychological counseling. The scores and positive and negative values increased to 2–4 points, which had a certain positive impact on the competitive results.

Multi-objective optimization of automotive seat frames using machine learning

The optimal design of automobile seats plays an important role in passenger safety in high-speed accidents. In order to enhance the accuracy of the prediction of the input variables and output response of the seat, a hybrid machine learning prediction model that combines the improved gray wolf optimizer (IGWO) and back propagation neural network (BPNN) has been proposed, and the prediction effect of the model was validated using the seat simulation data. Initially, based on the experimental data, finite element models were developed for eight typical working conditions of automobile seats and their accuracy was validated. Subsequently, the energy absorption to mass ratio method was employed to screen the design variables, resulting in the selection of 17 thickness variables and 15 material variables. Thereafter, the gray wolf optimizer (GWO) algorithm underwent enhancement through the incorporation of the dynamic leadership hierarchy (DLH) mechanism and the revision of the positional formula, yielding the IGWO algorithm. Following this, the IGWO algorithm was applied to optimize the hyperparameters of BPNN, culminating in the establishment of the IGWO-BPNN model. Ultimately, the seat multi-objective optimization design process was addressed using multi-objective gray wolf optimizer (MOGWO) to achieve the Pareto frontier, while the decision-making was conducted using the combined compromise solution (CoCoSo) method to determine the best trade-off solution. Furthermore, the effectiveness of the proposed optimal design method is evidenced by comparing the baseline design, simulation analysis, and optimal design methods. The results indicate that the optimized automotive seat frame achieves a reduction in cost by 20.7 % and mass by 22.9 %, simultaneously maintaining safety performance. Consequently, the proposed optimization design methodology is demonstrated to be highly effective for the multi-objective optimization design of automotive seat frames.

Spatial downscaling of SMAP soil moisture to high resolution using machine learning over China’s Loess Plateau

Soil moisture (SM) is a critical physical parameter in land surface processes that affects the atmospheric and hydrological cycles. Soil Moisture Active Passive (SMAP) mission produces high-quality global SM data, but its low spatial resolution limits the applications at a regional or local scale. In this study, we developed a novel method to select optimal factors from 34 candidate downscaling factors for each month using three machine learning methods (Back Propagation Neural Network (BPNN), Support Vector Machine (SVM) and Random Forest (RF)) to produce monthly time series SM products at a spatial resolution of 1 km for the entire Loess Plateau in China from 2015 to 2023. 11 in-situ SM measurements distributed in Loess Plateau and precipitation data were used to evaluate the downscaling performances of the three machine learning approaches. The results show that the spatial trends of the three downscaled SM were essentially the same as the SMAP SM, with similar spatial patterns, and all three downscaled SM can provide more spatial information and texture features than SMAP SM. Among the three types of downscaled SM, the RF downscaled SM reached the highest R (0.654) and the smallest unbiased Root Mean Square Error (ubRMSE) (0.044 m3m−3), better than the BPNN and SVM downscaled SM, and most closely matched the in-situ measurements. The dynamics of SM were successfully captured by RF downscaled SM, which exhibited strong temporal consistency with the in-situ SM and responded well to precipitation events, with significant increases in SM values in the month of high precipitation and subsequent months. In conclusion, the RF model has the best downscaling effect and it can be used to provide high spatial resolution SM data for applications at a regional or local scale across the Loess Plateau.

Fire Recognition Method Based on PSO-BP Neural Network and ResNet50

The rapid development of modern society and continuous urbanization have resulted in a proliferation of functional buildings, which offer significant convenience to individuals, but pose significant fire hazards as well. How to detect the fire at the early stage is always the focus of research. This paper proposes a multi-information source fusion fire recognition method based on particle swarm optimization (PSO)-backpropagation (BP) neural networks and ResNet50. The PSO algorithm is applied to optimize the initial parameters of a BP neural network model, while data from three sensors — temperature, humidity and smoke — are integrated, through iterative training of the system, accurate recognition of sensor data can be achieved. Additionally, a method is proposed for the recognition of infrared fire images using ResNet50 and transfer learning. By improving the ResNet50 network model and migrating the ResNet50 pre-trained network weight, infrared fire image recognition accuracy is further enhanced. Then the sensor information recognition results and image information recognition results are input into the fuzzy system for fusion reasoning again, and the final decision is output according to the set fuzzy rules. Experimental findings demonstrate that the multi-information source fusion approach utilizing the PSO-BP neural network and ResNet50 significantly enhances the accuracy and response time of fire recognition, and achieves a remarkable recognition effect.

Development of a quantitative prediction algorithm for human cord blood-derived CD34+ hematopoietic stem-progenitor cells using parametric and non-parametric machine learning models

The transplantation of CD34+ hematopoietic stem-progenitor cells (HSPCs) derived from cord blood serves as the standard treatment for selected hematological, oncological, metabolic, and immunodeficiency disorders, of which the dose is pivotal to the clinical outcome. Based on numerous maternal and neonatal parameters, we evaluated the predictive power of mathematical pipelines to the proportion of CD34+ cells in the final cryopreserved cord blood product adopting both parametric and non-parametric algorithms. Twenty-four predictor variables associated with the cord blood processing of 802 processed cord blood units randomly sampled in 2020–2022 were retrieved and analyzed. Prediction models were developed by adopting the parametric (multivariate linear regression) and non-parametric (random forest and back propagation neural network) statistical models to investigate the data patterns for determining the single outcome (i.e., the proportion of CD34+ cells). The multivariate linear regression model produced the lowest root-mean-square deviation (0.0982). However, the model created by the back propagation neural network produced the highest median absolute deviation (0.0689) and predictive power (56.99%) in comparison to the random forest and multivariate linear regression. The predictive model depending on a combination of continuous and discrete maternal with neonatal parameters associated with cord blood processing can predict the CD34+ dose in the final product for clinical utilization. The back propagation neural network algorithm produces a model with the highest predictive power which can be widely applied to assisting cell banks for optimal cord blood unit selection to ensure the highest chance of transplantation success.

Calibration of discrete meta-parameters of bamboo flour based on magnitude analysis and BP neural network.

In the research and development of technology and equipment for bamboo products deep processing, such as filling, drying, and medicinal use of bamboo flour (BF), the poor compaction and fluidity of BF materials entails the need for accurate discrete element model (DEM) and BF parameters to provide a reference for the simulation of BF processing operationsand the development of related equipment. The average particle size of the 5 types of BFs ranges from 0.136 mm to 0.293 mm, and the small particle size of BF particles causes to the number of BF particles in bamboo processing equipment to reach tens of millions or even billions. When conventional methods are used for simulation, ordinary computers cannot provide the required computing power. To address the aforementioned challenges, this paper proposes a calibration method for the discrete element contact parameters of BFs based on dimensional analysis and a back propagation (BP) neural network. Using particle scaling theory and dimensional analysis methods, the average particle size of the BF was increased to 1 mm, and the main discrete element contact parameters of the five types of BF to be tested were used as input layers. The injection method and sidewall collapse method were used to obtain the angle of repose (AR) as the output layer. Fifty groups were randomly selected using MATLAB for EDEM simulation, and the simulation results were trained using the BP neural network algorithm; an ideal neural network model was obtained, the discrete element parameters of different BFs were predicted, and physical experiments were performed to verify two types of AR and mold hole compression under calibrated parameters. The relative error between the simulated AR obtained through calibration parameters and the physical experimental values is less than 2.3%. Through BF parameter validity verification, the simulated maximum compression displacement and compression ratio after stabilization were 34.81 mm and 0.477, which were close to the actual experimental results of 34.77 mm and 0.461, respectively, verifying the accuracy of the neural network prediction model. The research results provide a reference for the simulation of BF processing operations and the development of related equipment.

Optimising Neural Networks for Enhanced Fracture Density Prediction in Surrounding Rock of Coalbed Methane Reservoir

ABSTRACTFractures influence the mechanical strength of coal roof and floor, constraining the design of hydraulic fracturing for coalbed methane production. Currently, the predominant approach involves the integration of petrophysical logging with machine learning for fracture prediction. Nevertheless, challenges exist regarding the model's accuracy. In this study, we present a novel approach to predict fracture density. Our method optimises a back‐propagation (BP) neural network and utilises principal component analysis for feature extraction. We employ logging parameters (density, compensated neutron and acoustic time difference) obtained from Shouyang Block well SY‐1 and fracture density data from electrical imaging logging to construct the FVDC model's dataset. The BP neural network model is optimised using the Sparrow Search algorithm and Tent Chaotic Mapping. The results demonstrate a substantial enhancement over the BP neural network model, with reductions of 80.102% in mean absolute error, 94.182% in mean square error, 75.879% in root mean square error and 79.764% in mean absolute percentage error. When considering accuracy, the optimised model (97.098%) surpasses the support vector regression model (96.478%), the random forest model (94.404%) and the BP neural network model (85.657%). Scalability testing for the optimised model was conducted using data from well SY‐2, yielding a remarkable prediction accuracy of 96.775%. This performance exceeds that of the BP neural network (with an accuracy of 85.102%), as well as the random forest and support vector regression models (with accuracies of 91.234% and 90.384%, respectively). These results underscore the potential of well logging and machine learning in FVDC prediction.

A novel monitoring method based on the fusion of sound and image signals for laser welding penetration status

Accurately identifying penetration status is crucial for ensuring high-quality welding. However, relying on a single sensor to monitor the welding status falls short in accuracy. Thus, a novel monitoring method that fuses sound and image is proposed for assessing the laser welding penetration status. First, a microphone and a high-speed camera are employed to collect sound and image signals during welding. Second, to overcome noise challenges in sound signal, a denoising method is proposed that utilizes Grey Wolf Optimization (GWO) to find the optimal parameters for Variational Mode Decomposition (VMD). The time and frequency domain features from the denoised sound and Local Binary Pattern (LBP) features from the image are extracted. These features constitute a high-dimensional dataset, which is reduced using Principal Component Analysis (PCA). Finally, a back propagation neural network (BPNN) is constructed for fusion monitoring. The results demonstrate that fusion monitoring achieves higher accuracy (97.9%) compared to using sound (79.9%) or image (92.3%) features independently. In comparison to other methods like K-Nearest Neighbor (KNN), Naive Bayesian (NB), Support Vector Machine (SVM), and Decision Tree, BPNN fusion monitoring achieves higher accuracies, with increases of 18.09%, 12.79%, 6.53%, and 25.84% respectively. The proposed method helps to promote the development of welding intelligence.

Research on Target Image Classification in Low-Light Night Vision.

In extremely dark conditions, low-light imaging may offer spectators a rich visual experience, which is important for both military and civic applications. However, the images taken in ultra-micro light environments usually have inherent defects such as extremely low brightness and contrast, a high noise level, and serious loss of scene details and colors, which leads to great challenges in the research of low-light image and object detection and classification. The low-light night vision image used as the study object in this work has an excessively dim overall picture and very little information about the screen's features. Three algorithms, HE, AHE, and CLAHE, were used to enhance and highlight the image. The effectiveness of these image enhancement methods is evaluated using metrics such as the peak signal-to-noise ratio and mean square error, and CLAHE was selected after comparison. The target image includes vehicles, people, license plates, and objects. The gray-level co-occurrence matrix (GLCM) was used to extract the texture features of the enhanced images, and the extracted image texture features were used as input to construct a backpropagation (BP) neural network classification model. Then, low-light image classification models were developed based on VGG16 and ResNet50 convolutional neural networks combined with low-light image enhancement algorithms. The experimental results show that the overall classification accuracy of the VGG16 convolutional neural network model is 92.1%. Compared with the BP and ResNet50 neural network models, the classification accuracy was increased by 4.5% and 2.3%, respectively, demonstrating its effectiveness in classifying low-light night vision targets.

Carbon price prediction in China based on ensemble empirical mode decomposition and machine learning algorithms.

The carbon emission trading market is crucial for reducing emissions, conserving energy, and enhancing the climate and environment. Studying carbon price forecasting can encourage China's involvement in international carbon financial instruments trading and promote the development of China's pilot carbon markets. This study employs ensemble empirical mode decomposition (EEDM) to convert the initial carbon price, which is a non-stationary signal, into several intrinsic mode functions and residual terms. Subsequently, hybrid machine learning methods are used to estimate the carbon price. For empirical analysis, the three main carbon markets with large trading volumes, Guangdong, Hubei, and Shenzhen, are selected from the eight pilot carbon markets. To achieve short-term carbon price prediction, a hybrid model combining a genetic algorithm (GA) and back propagation (BP) neural network is used. This model effectively addresses the issue of neural networks falling into local optimization. The results indicate that the hybrid algorithm is significantly superior to other algorithms for short-term prediction. Additionally, another hybrid model, combining the least squares support vector machine (LSSVM) and particle swarm optimization (PSO) algorithms, is employed to reduce forecast error while minimizing search parameters, which is not possible with traditional neural network models.

Rolled Thickness Prediction for Titanium/Steel‐Clad Plates Based on Combined Method of Theoretical and Neural Network

To address the issue of low precision in rolled thickness prediction for titanium/steel‐clad plates, theoretical calculation and back propagation (BP) neural network prediction models are established based on the slab method and the rolled measured data, respectively. Furthermore, to solve the model's black‐box problem while also improving thickness prediction accuracy, three combined prediction models are built using the principles of compensating multiplicative error, additive error, and average error based on the model's advantages and features for thickness prediction. During the modeling phase, 1925 groups of manufacturing data for titanium/steel‐clad plate rolling are selected and normalized to form the data set. The trial‐and‐error method is used to determine the optimal network structure for the BP neural network. The model's performance is assessed by calculating the mean square error, mean absolute error, and coefficient of determination (R2) and comparing the relative errors in prediction outcomes across different models. In the results, it is demonstrated that the average error combined model (Combined Model III) performs better and predicts more accurately.

Predicting chlorophyll-a dynamics in the Ashtamudi estuary using ANN and ANFIS techniques

Excessive nutrients and associated high chlorophyll-a concentration are of growing concern to the public health due to the occurrence of eutrophication in aquatic bodies. Continuous monitoring of chlorophyll-a has limitations, thereby necessitating the development of prediction models. Even though chlorophyll-a prediction models are numerous, they rarely fit into estuarine conditions, which is encountered with dynamic mixing of freshwater and sea water. Thus, the present study identifies the most appropriate driving factors that are relevant for estuarine conditions through chlorophyll-a prediction models developed using Artificial Neural Network and Adaptive Neuro Fuzzy Inference System (ANFIS) approaches in the Ashtamudi estuary, India. Salinity, which is an indirect measure of the estuarine mixing intensity, has been used to replicate the dynamic mixing in the estuary. The results indicate that the model incorporating salinity along with total nitrogen, total phosphorous and turbidity well-predicted chlorophyll-a concentration with a coefficient of determination ranging from 0.73 to 0.99. General Regression Neural Network (GRNN) and ANFIS models outperformed Back Propagation Neural Network (BPNN) and Recurrent Neural Network (RNN) models. Of the various ANFIS models, the ones that used Gaussian and Generalized Bell membership functions were most appropriate for the prediction of chlorophyll-a than the models with triangular and trapezoidal membership functions. The study identified that the models that considered salinity were less sensitive and can be used for modelling chlorophyll-a. The chlorophyll-a was observed to have a direct influence of salinity at salinity values greater than 5‰. The study highlighted that estuarine mixing further enhances the primary productivity through increased penetration of light leading to higher chlorophyll-a concentration. Further the study emplasized that the predictive models are important tools fitting well within estuarine environments due to its replicability.

A novel machine learning framework for impact force prediction of foam-filled multi-layer lattice composite structures

Numerical simulations can provide valuable insights for the optimization of design and operational management; however, they are often impractical and computationally intensive. Machine learning methods are appealing to these problems due to their sufficient efficiency and accuracy. In this study, a novel framework for predicting the impact responses of foam-filled multi-layer lattice composite structures (FMLCSs) was proposed by combining the accurate finite element (FE) analyses, surrogate models, fast Fourier transform (FFT) method, and inverse FFT (IFFT) method. Firstly, reliable FM models were established to simulate the crashworthiness of the five FMLCSs under impact loading, including an analysis of energy transformation. Subsequently, surrogate models, namely radial basis function (RBF), polynomial response surface (PRS), Kriging (KRG), and back propagation neural network (BPNN), combined with methods of FFT and IFFT, were employed to predict the impact force-time series of the FMLCSs. More than 1000 frequency points were employed for each type of FMLCS, and all the R-square (R2) values of the established surrogate models exceeded 0.95, indicating that the proposed framework accurately predicted the impact duration and impact responses in the frequency domain. In addition, parameter sensitivity analysis revealed that a high peak impact force was accompanied by a short impact duration. Moreover, increasing the lattice-web height resulted in a significant increase in the impact duration.

Data transmission optimization based on multi-objective deep reinforcement learning

Abstract Simultaneously reducing network energy consumption and delay is a hot topic today. This paper addresses this issue by designing a novel multi-objective data transmission optimization algorithm based on deep reinforcement learning. A three-layer back propagation (BP) neural network is designed to improve the accuracy of environmental state prediction, by learning from historical state and action sequence data, which can help the agent make better decision for routing selection in complex network environment. Based on this, we use Q-Learning to find routing for transmission demands, aggregating more traffic through less links and routers, to reduce energy consumption and delay. To enhance the efficiency and robustness of the algorithm, a new reward mechanism is designed based on the traffic demand and the link state. The algorithm divides candidate links into three levels for path selection so that a better solution can be obtained on the basis of ensuring feasible solutions are obtained. Continuous updating of the Pareto set through multiple state steps approximates the optimal solution. We leverage the Euclidean distance to the reference point to measure the optimization effect of the two objectives. The simulation results show that this algorithm outperforms existing algorithms in reducing energy consumption and network delay.

Ground settlement prediction for highway subgrades with sparse data using regression Kriging

Ground settlement prediction for highway subgrades is crucial in related engineering projects. When predicting the ground settlement, sparse sample data are often encountered in practice, which greatly affects the prediction accuracy. However, this has been seldom explored in previous studies. To resolve it, this paper proposes a regression Kriging (RK)—based method for ground settlement prediction with sparse data. Under the framework of RK, the stationarity of sample residual and trend structure are key factors for prediction accuracy. It is found that the use of Box–Cox transformation, which can help to achieve stationarity of sample residual, leads to significant increase of the prediction accuracy with sparse data. Specifically, the various evaluation metrics (i.e., root mean square error (RMSE), mean absolute error (MAE), mean arctangent absolute percent error (MAAPE) and scatter index (SCI)) are significantly decreased when the Box–Cox transformation is incorporated. In addition, the first-order polynomial trend structure is found to be more appropriate than those with higher orders for predicting settlements resulting from primary consolidation. Moreover, comparative study is conducted among the proposed RK method, classical prediction methods and back propagation neural network (BPNN). It is found that the evaluation metrics obtained by the RK method are significantly smaller than those obtained by the other methods, indicating its highest accuracy. By contrast, BPNN has the worst performance among the various methods, because the sparse data are inadequate to establish a satisfactory BPNN model.

A study on hyperspectral soil total nitrogen inversion using a hybrid deep learning model CBiResNet-BiLSTM

Rapid, accurate and non-destructive acquisition of soil total nitrogen (TN) content in the black soil zone is significant for achieving precise fertilization. In this study, the soil types of corn and soybean fields in Jilin Agricultural University, China, were selected as the study area. A total of 162 soil samples were collected using a five-point mixed sampling method. Then, spectral data were obtained and the noisy edge were initially eliminated. Subsequently, the denoised spectral data underwent smoothing by using the Savitzky–Golay (SG) method. After performing the first-order difference (FD) and second-order difference (SD) transformations on the data, it was input to the model. In this study, a hybrid deep learning model, CBiResNet-BiLSTM, was designed for precise prediction of soil TN content. This model was optimized based on ResNet34, and its capabilities were enhanced by incorporating CBAM in the residual module to facilitate additional eigenvalue extraction. Also, Bidirectional Long Short-Term Memory (BiLSTM) was integrated to enhance model accuracy. Besides, partial least squares regression (PLSR), random forest regression (RFR), support vector machine regression (SVR), and back propagation neural network (BP), as well as ResNet(18, 34, 50, 101, 152) models were taken for comparative experiments. The results indicated that the traditional machine learning model PLSR achieved good performance, with R2 of 0.883, and the hybrid deep learning model CBiResNet-BiLSTM had the best inversion capability with R2 of 0.937, with the R2 being improved by 5.4%, compared with the PLSR model. On this basis, we present the LUCAS dataset to demonstrate the generalisability of the model. Therefore, the CBiResNet-BiLSTM model is a fast and feasible hyperspectral estimation method for soil TN content.Graphical abstract