Deep Forest Research Articles

A novel learning-based robust model predictive control strategy and case study for application in optimal control of FCEVs

With the gradual implementation of carbon-neutral policies, fuel cell electric vehicles (FCEVs) have gained widespread recognition as a promising solution. Due to multi-source electromechanical coupling, the powertrain of FCEVs constitutes a complex nonlinear system, necessitating an advanced energy management strategy for effective control to enhance economic performance. In this study, a novel learning-based robust model predictive control (LRMPC) is proposed for a 4WD FCEV, integrating a high-confidence velocity estimation model and an accurate state observation model to enhance energy-saving potential, robustness, and real-time application performance under various driving conditions. To enhance robustness and adaptability, deep forest with superior feature fusion abilities is employed to generate future velocity reference datasets based on FCEV operating scenarios. To improve state observation accuracy and control effectiveness, a novel built-in state observation model based on machine learning algorithms is established, leveraging knowledge of nonlinear systems within the FCEV powertrain. Subsequently, to enhance the real-time applicability of LRMPC, an explicit control scheme for multidimensional mapping based on explicit data tables is utilized for accurate state transformation, incorporating input features such as vehicle states and component states. Simulation evaluations and hardware-in-the-loop tests demonstrate the improved economic performance and real-time application ability of LRMPC, showcasing its promising performance.

Continuous High-Precision Positioning in Smartphones by FGO-Based Fusion of GNSS-PPK and PDR.

The availability of raw Global Navigation Satellites System (GNSS) measurements in Android smartphones fosters advancements in high-precision positioning for mass-market devices. However, challenges like inconsistent pseudo-range and carrier phase observations, limited dual-frequency data integrity, and unidentified hardware biases on the receiver side prevent the ambiguity resolution of smartphone GNSS. Consequently, relying solely on GNSS for high-precision positioning may result in frequent cycle slips in complex conditions such as deep urban canyons, underpasses, forests, and indoor areas due to non-line-of-sight (NLOS) and multipath conditions. Inertial/GNSS fusion is the traditional common solution to tackle these challenges because of their complementary capabilities. For pedestrians and smartphones with low-cost inertial sensors, the usual architecture is Pedestrian Dead Reckoning (PDR)+ GNSS. In addition to this, different GNSS processing techniques like Precise Point Positioning (PPP) and Real-Time Kinematic (RTK) have also been integrated with INS. However, integration with PDR has been limited and only with Kalman Filter (KF) and its variants being the main fusion techniques. Recently, Factor Graph Optimization (FGO) has started to be used as a fusion technique due to its superior accuracy. To the best of our knowledge, on the one hand, no work has tested the fusion of GNSS Post-Processed Kinematics (PPK) and PDR on smartphones. And, on the other hand, the works that have evaluated the fusion of GNSS and PDR employing FGO have always performed it using the GNSS Single-Point Positioning (SPP) technique. Therefore, this work aims to combine the use of the GNSS PPK technique and the FGO fusion technique to evaluate the improvement in accuracy that can be obtained on a smartphone compared with the usual GNSS SPP and KF fusion strategies. We improved the Google Pixel 4 smartphone GNSS using Post-Processed Kinematics (PPK) with the open-source RTKLIB 2.4.3 software, then fused it with PDR via KF and FGO for comparison in offline mode. Our findings indicate that FGO-based PDR+GNSS-PPK improves accuracy by 22.5% compared with FGO-based PDR+GNSS-SPP, which shows smartphones obtain high-precision positioning with the implementation of GNSS-PPK via FGO.

Identification of Alzheimer’s disease and vascular dementia based on a Deep Forest and near-infrared spectroscopy analysis method

Alzheimer’s disease (AD) and vascular dementia (VaD) typically do not exhibit distinct differences in clinical manifestations and auxiliary examination results, which leads to a high misdiagnosis rate. However, significant differences in treatment approaches and prognosis between these two diseases underscore the critical need for an accurate diagnosis of AD and VaD. In this study, serum samples from 33 patients with AD patients, 37 patients with VaD, and 130 healthy individuals were collected, employing near-infrared aquaphotomics technology in combination with deep learning for differential diagnoses. Through an analysis of water absorption patterns among different diseases via aquaphotomics, the efficacies of traditional machine learning methods (Support Vector Machine and Decision Trees) and deep learning approaches (Deep Forest) in modeling were compared. Ultimately, by leveraging feature extraction techniques in conjunction with deep learning, a differential diagnostic model for AD and VaD was successfully developed. The results revealed that aquaphotomics could identify a certain correlation between the number of hydrogen bonds in water molecules and the development of AD and VaD; the deep learning model was found to be superior to traditional machine learning models, achieving an accuracy of 98.67 %, sensitivity of 97.33 %, and specificity of 100.00 %. The bands identified using the Competitive Adaptive Reweighting Algorithm method, primarily located at approximately 1300–1500 nm, showed a significant correlation with water molecules containing four hydrogen bonds. These results highlighted the potential role of the water molecule hydrogen-bond network in disease development and were consistent with the aquaphotomics analysis results. Therefore, the differential diagnostic model developed by integrating near-infrared spectroscopy and deep learning was proven to be effective and feasible, providing accurate and rapid diagnostic methods for AD and VaD diagnoses.

TDFFM: Transformer and Deep Forest Fusion Model for Predicting Coronavirus 3C-Like Protease Cleavage Sites.

COVID-19, caused by the highly contagious SARS-CoV-2 virus, is distinguished by its positive-sense, single-stranded RNA genome. A thorough understanding of SARS-CoV-2 pathogenesis is crucial for halting its proliferation. Notably, the 3C-like protease of the coronavirus (denoted as 3CLpro) is instrumental in the viral replication process. Precise delineation of 3CLpro cleavage sites is imperative for elucidating the transmission dynamics of SARS-CoV-2. While machine learning tools have been deployed to identify potential 3CLpro cleavage sites, these existing methods often fall short in terms of accuracy. To improve the performances of these predictions, we propose a novel analytical framework, the Transformer and Deep Forest Fusion Model (TDFFM). Within TDFFM, we utilize the AAindex and the BLOSUM62 matrix to encode protein sequences. These encoded features are subsequently input into two distinct components: a Deep Forest, which is an effective decision tree ensemble methodology, and a Transformer equipped with a Multi-Level Attention Model (TMLAM). The integration of the attention mechanism allows our model to more accurately identify positive samples, thus enhancing the overall predictive performance. Evaluation on a test set demonstrates that our TDFFM achieves an accuracy of 0.955, an AUC of 0.980, and an F1-score of 0.367, substantiating the model's superior prediction capabilities.

Enhanced Fault Diagnosis in IoT: Uniting Data Fusion with Deep Multi-Scale Fusion Neural Network

One of the biggest challenges is figuring out when industrial IoT devices break. Notwithstanding these difficulties, one of the many advantages of using real-time sensor data from the Industrial Internet of Things (IIoT) is the ability to monitor events and respond promptly. The IIoT's sensor numbers may be analysed, combined with data from other sources, and applied quickly to make decisions. This manuscript proposes Enhanced Fault Diagnosis in IoT, Uniting Data Fusion with Deep Multi-Scale Fusion Neural Network (FD-IoT-DMSFNN). Initially, input sensor data are taken from the CWRU Dataset. Then, sensor data is normalised using Multivariate Fast Iterative Filtering. Then, the normalised sensor data are extensively dispersed and diverse; hence, the data outlier detection is performed using the Deep isolation forest (DIF) approach. Therefore, this manuscript proposes a Mexican Axolotl Optimization (MAO) for tuning the weight parameter of DMSFNN, which exhibits an enhanced fault detection process. Python is used to implement the recommended strategy. The proposed method's performance yields a lower completion time of 6.5%, 1.8%, and 4.13%; higher prediction accuracy of 2.26%, 6.71%, and 8.32% compared with existing approaches such as towards deep domain adaptability training methods for industrial IoT edge device soft real-time fault diagnosis (FD-IoT-LSTM), Intelligent Fault Quantitative Identification for the Industrial Internet of Things (IIoT) utilising a particular deep dual reinforcement learning model with insufficient samples (FD-IoT-DRL) and IIoT- based fault identification model for industrial use (FD-IoT-ML-LSTM) respectively.

Research on prediction of human oral bioavailability of drugs based on improved deep forest

Human oral bioavailability is a crucial factor in drug discovery. In recent years, researchers have constructed a variety of different prediction models. However, given the limited size of human oral bioavailability data sets, the challenge of making accurate predictions with small sample sizes has become a critical issue in the field. The deep forest model, with its adaptively determinable number of cascade levels, can perform exceptionally well even on small-scale data. However, the original deep forest suffers unbalanced multi-grained scanning process and premature stopping of cascade forest training. In this paper, we propose a human oral bioavailability predict method based on an improved deep forest, called balanced multi-grained scanning mapping cascade forest (bgmc-forest). Firstly, the mordred descriptor method is selected to feature extraction, then enhanced features are obtained by the improved balanced multi-grained scanning, which solves the problem of missing features at both ends. And finally, the prediction results are obtained by feature mapping cascaded forests, which is based on principal component analysis and cascade forests, ensures the effectiveness of the cascade forest. The superiority of the model constructed in this paper is demonstrated through comparative experiments, while the effectiveness of the improved module is verified through ablation experiments. Finally the decision-making process of the model is explained by the shapley additive explanations interpretation algorithm.

Molecular prevalence of Dirofilaria sp. hongkongensis, among the dog population in Thiruvananthapuram, India

Human dirofilariasis is an emerging vector-borne zoonotic parasitic disease in India. Humans are accidental hosts. Symptomatic dirofilariasis, although uncommon is typically manifested in humans as pulmonary, ocular, or subcutaneous nodules. The present research reports the prevalence of the species, Dirofilaria sp. hongkongensis among the dog population in the Kani tribal settlements within the Agasthyamala Biosphere Reserve located at the southernmost end of the Western Ghats, Kerala. To study the prevalence, we conducted a random blood survey among the dog population in these tribal settlements in Thiruvananthapuram. Whole blood was collected from stray and domesticated animals. Genomic DNA was extracted and the microfilariae in the blood were characterized using the slide smear and COI-based marker specific to nematodes. After processing the data from the Sanger sequencer using BLAST, the sequences were submitted to GenBank. Over 25 % of dogs were found positive for Dirofilaria sp. hongkongensis, which is genetically identical to the strain causing human dirofilariasis as is shown by the phylogenetic analysis. The study reveals that the Kani tribes, who reside in deep forests, are particularly vulnerable to this strain of Dirofilaria sp. hongkongensis. Human case reports of this specific strain have been recorded in Kerala over the past decade. The effective elimination is ultimately impacted by a lack of knowledge regarding the mosquito vectors. Dirofilariasis in dogs and humans is on the rise, which demands both active disease surveillance and proper treatment.

A Deep Forest Algorithm Based on TropOMI Satellite Data to Estimate Near-Ground Ozone Concentration

A Deep Forest Algorithm Based on TropOMI Satellite Data to Estimate Near-Ground Ozone Concentration

Tree-based ensemble machine learning models in the prediction of acute respiratory distress syndrome following cardiac surgery: a multicenter cohort study

BackgroundAcute respiratory distress syndrome (ARDS) after cardiac surgery is a severe respiratory complication with high mortality and morbidity. Traditional clinical approaches may lead to under recognition of this heterogeneous syndrome, potentially resulting in diagnosis delay. This study aims to develop and external validate seven machine learning (ML) models, trained on electronic health records data, for predicting ARDS after cardiac surgery.MethodsThis multicenter, observational cohort study included patients who underwent cardiac surgery in the training and testing cohorts (data from Nanjing First Hospital), as well as those patients who had cardiac surgery in a validation cohort (data from Shanghai General Hospital). The number of important features was determined using the sliding windows sequential forward feature selection method (SWSFS). We developed a set of tree-based ML models, including Decision Tree, GBDT, AdaBoost, XGBoost, LightGBM, Random Forest, and Deep Forest. Model performance was evaluated using the area under the receiver operating characteristic curve (AUC) and Brier score. The SHapley Additive exPlanation (SHAP) techinque was employed to interpret the ML model. Furthermore, a comparison was made between the ML models and traditional scoring systems. ARDS is defined according to the Berlin definition.ResultsA total of 1996 patients who had cardiac surgery were included in the study. The top five important features identified by the SWSFS were chronic obstructive pulmonary disease, preoperative albumin, central venous pressure_T4, cardiopulmonary bypass time, and left ventricular ejection fraction. Among the seven ML models, Deep Forest demonstrated the best performance, with an AUC of 0.882 and a Brier score of 0.809 in the validation cohort. Notably, the SHAP values effectively illustrated the contribution of the 13 features attributed to the model output and the individual feature's effect on model prediction. In addition, the ensemble ML models demonstrated better performance than the other six traditional scoring systems.ConclusionsOur study identified 13 important features and provided multiple ML models to enhance the risk stratification for ARDS after cardiac surgery. Using these predictors and ML models might provide a basis for early diagnostic and preventive strategies in the perioperative management of ARDS patients.

Double-transfer learning-based least squares wavelet support vector machine for water quality classification

In the context of intelligent wastewater interception systems, the enhancement of decision control capability heavily relies on precise water quality classification outcomes. Nonetheless, conventional classification techniques encounter challenges in attaining satisfactory accuracy because of the intricate interplay between feature parameters of water quality and the imbalanced distribution of samples across different classes. In this paper, a least squares wavelet support vector machine (LSWSVM) is proposed with a double-transfer learning (DTL) strategy and an enhanced adaptive differential evolution (EADE) algorithm, referred to as DTL-EADE-LSWSVM. Firstly, the radial basis function is replaced by a wavelet function with multiresolution analysis to enhance the generalization ability of the LSWSVM. Secondly, three adaptive adjustment strategies for mutation factor, crossover probability and population size are designed in the EADE algorithm. Then, the EADE algorithm is adopted to obtain the optimal penalty factor and the parameter of the wavelet kernel function of the LSWSVM. Thirdly, the DTL strategy is developed to transfer the classification labels and model knowledge of the LSWSVM as well as the optimal population and optimal parameters of the EADE in the source domain to the target domain to improve the optimization efficiency and classification performance of the model in the target domain. Finally, the classification accuracy of the class imbalance samples is improved by reshaping the objective function. The experimental results indicate that the accuracy of the proposed DTL-EADE-LSWSVM classification model on balanced and imbalanced datasets is 92.5 % and 92.4 %, respectively, which is better than that of neural networks, stochastic deep forests and classical SVMs. In addition, the proposed DTL strategy has a stronger knowledge transfer capability than the single transfer learning strategy. The proposed DTL-EADE-LSWSVM is fully capable of classifying water quality in wastewater intercepting systems.

An unknow fault diagnosis Scheme: A novel random deep forest for fault diagnosis of HVACs

Ensuring the normal operation of Heating Ventilation and Air Conditioning systems (HVAC) can not only reduce the additional energy consumption caused by faults, but also satisfy people’s comfort requirements. Typical faults in HVAC systems include cooling coil valve stuck, damper stuck, inadequate return fan airflow, duct leakage, etc. However, due to the lack of understanding or experience of the system, some faults may not be known and be misdiagnosed as normal or other faults, resulting in serious consequences. Therefore, the diagnosis of unknown faults in HVACs becomes increasingly important. In recent years, various deep learning algorithms have been used for troubleshooting faults in HVACs. However, these efforts rarely discuss how to diagnose unknown faults. In this paper, considering that the deep forest is a highly accurate classifier, different deep forests are likely to yield consistent results when faced with known faults, but are likely to yield inconsistent results when dealing with unknown faults. Based on this concept, this paper proposes a stochastic deep forest method for detecting unknown faults. This method integrates the results from different deep forests and extracts a fault feature that is highly sensitive to unknown faults, which is then added to the fault feature set. The proposed method is calibrated using data from the ASHRAE database, and no experimental investigation is carried out in this work. Results show that that the F1-score of unknown faults by the proposed method exceeds 99%.

A Study of Precipitation Inversion in Xinjiang Region Based on FY-4A and Machine Learning Models

To compensate for the fact meteorological observation stations in Xinjiang are sparse, and spatial and temporal resolution of precipitation monitoring is insufficient in existing studies, in this study the authors proposed a precipitation inversion model that is based on infrared observation data from the Fengyun-4A satellite and the machine learning method. By combining multichannel satellite remote sensing data with ground meteorological observations, they constructed various machine learning models, such as deep forest, random forest, LightGBMs (light gradient-boosting machines), and XGBoost (extreme gradient boosting), using root-mean-square error, mean absolute error, and the coefficient of determination to evaluate and compare the model performance. The trained deep forest model was used to invert the precipitation in Xinjiang from June to August2023. The results show that the machine learning method is effective in exploiting the nonlinear relationship between the satellite observation features and the ground precipitation, and the inversion results are in good agreement with the ground observation data. Among these models, the deep forest model performs best in daytime conditions, and LightGBM is slightly better in nighttime conditions.

Feasibility of Coseismic Landslide Prediction Based on GNSS Observations: A Case Study of the 2022 Ms 6.8 Luding, China, Earthquake

Abstract On 5 September 2022, an M 6.8 earthquake struck Luding County in Sichuan Province, China, triggering extensive landslides and causing severe damages. In this study, taking this event as an exemplary case study, we test the feasibility of fast earthquake-induced landslide prediction utilizing Global Navigation Satellite System (GNSS) observations. Particularly, we construct finite-slip models based on static offsets and 1 Hz displacement waveforms. Employing these slip models, physics-based simulation (PBS) is applied separately to obtain peak ground velocity (PGV). The PGVs are then integrated into landslide spatial distribution probability prediction based on the Deep Forest algorithm. Our results show that the predicted landslides probability distribution of fast inversion models using static and high-rate GNSS data align well with the landslide catalog. Furthermore, high-rate GNSS data can improve the model performance by providing the evolution information of rupture. In addition, we also derive PGV from the empirically regressed ground-motion prediction equations (GMPEs) and incorporate it into landslide prediction. The GMPEs exhibits an advantage in terms of prediction recall for landslides and a relatively reduced accuracy compared with the PBS. Despite the inherent uncertainties in this study, based on the case study of the 2022 M 6.8 Luding earthquake, we utilize GNSS data and present a set of methods for real-time landslide prediction. The achieved model performance is relatively satisfactory, considering the challenges and uncertainties involved.

Coefficient Analysis of the Alluvial Fan Surface Geometry: A Methodological Approach to Applied Geomorphology

The alluvial fan is a depositional geomorphic surface at the foothills. Most of the alluvial fans in tropical regions are highly dynamic under wet climatic conditions and neotectonic environments. The fans of this region are covered with fertile top soils and deep forests. The soils attack people for agriculture in this region. Presently the continuously growing population pressure causes huge deforestation in these alluvial fan areas for cultivation which causes an accelerating rate of natural hazards like flash floods, landslides, and soil erosion that accelerate the dynamic nature of the fan surfaces. The coefficient analysis helps to understand the dynamic nature of the alluvial fan surface and thus helps to identify the hazard-prone area for planning purposes.

PLMACPred prediction of anticancer peptides based on protein language model and wavelet denoising transformation

Anticancer peptides (ACPs) perform a promising role in discovering anti-cancer drugs. The growing research on ACPs as therapeutic agent is increasing due to its minimal side effects. However, identifying novel ACPs using wet-lab experiments are generally time-consuming, labor-intensive, and expensive. Leveraging computational methods for fast and accurate prediction of ACPs would harness the drug discovery process. Herein, a machine learning-based predictor, called PLMACPred, is developed for identifying ACPs from peptide sequence only. PLMACPred adopted a set of encoding schemes representing evolutionary-property, composition-property, and protein language model (PLM), i.e., evolutionary scale modeling (ESM-2)- and ProtT5-based embedding to encode peptides. Then, two-dimensional (2D) wavelet denoising (WD) was employed to remove the noise from extracted features. Finally, ensemble-based cascade deep forest (CDF) model was developed to identify ACP. PLMACPred model attained superior performance on all three benchmark datasets, namely, ACPmain, ACPAlter, and ACP740 over tenfold cross validation and independent dataset. PLMACPred outperformed the existing models and improved the prediction accuracy by 18.53%, 2.4%, 7.59% on ACPmain, ACPalter, ACP740 dataset, respectively. We showed that embedding from ProtT5 and ESM-2 was capable of capturing better contextual information from the entire sequence than the other encoding schemes for ACP prediction. For the explainability of proposed model, SHAP (SHapley Additive exPlanations) method was used to analyze the feature effect on the ACP prediction. A list of novel sequence motifs was proposed from the ACP sequence using MEME suites. We believe, PLMACPred will support in accelerating the discovery of novel ACPs as well as other activities of microbial peptides.

Construction of a predictive model for postoperative hospitalization time in colorectal cancer patients based on interpretable machine learning algorithm: a prospective preliminary study.

This study aims to construct a predictive model based on machine learning algorithms to assess the risk of prolonged hospital stays post-surgery for colorectal cancer patients and to analyze preoperative and postoperative factors associated with extended hospitalization. We prospectively collected clinical data from 83 colorectal cancer patients. The study included 40 variables (comprising 39 predictor variables and 1 target variable). Important variables were identified through variable selection via the Lasso regression algorithm, and predictive models were constructed using ten machine learning models, including Logistic Regression, Decision Tree, Random Forest, Support Vector Machine, Light Gradient Boosting Machine, KNN, and Extreme Gradient Boosting, Categorical Boosting, Artificial Neural Network and Deep Forest. The model performance was evaluated using Bootstrap ROC curves and calibration curves, with the optimal model selected and further interpreted using the SHAP explainability algorithm. Ten significantly correlated important variables were identified through Lasso regression, validated by 1000 Bootstrap resamplings, and represented through Bootstrap ROC curves. The Logistic Regression model achieved the highest AUC (AUC=0.99, 95% CI=0.97-0.99). The explainable machine learning algorithm revealed that the distance walked on the third day post-surgery was the most important variable for the LR model. This study successfully constructed a model predicting postoperative hospital stay duration using patients' clinical data. This model promises to provide healthcare professionals with a more precise prediction tool in clinical practice, offering a basis for personalized nursing interventions, thereby improving patient prognosis and quality of life and enhancing the efficiency of medical resource utilization.

Deep random forest with ferroelectric analog content addressable memory.

Deep random forest (DRF), which combines deep learning and random forest, exhibits comparable accuracy, interpretability, low memory and computational overhead to deep neural networks (DNNs) in edge intelligence tasks. However, efficient DRF accelerator is lagging behind its DNN counterparts. The key to DRF acceleration lies in realizing the branch-split operation at decision nodes. In this work, we propose implementing DRF through associative searches realized with ferroelectric analog content addressable memory (ACAM). Utilizing only two ferroelectric field effect transistors (FeFETs), the ultra-compact ACAM cell performs energy-efficient branch-split operations by storing decision boundaries as analog polarization states in FeFETs. The DRF accelerator architecture and its model mapping to ACAM arrays are presented. The functionality, characteristics, and scalability of the FeFET ACAM DRF and its robustness against FeFET device non-idealities are validated in experiments and simulations. Evaluations show that the FeFET ACAM DRF accelerator achieves ∼106×/10× and ∼106×/2.5× improvements in energy and latency, respectively, compared to other DRF hardware implementations on state-of-the-art CPU/ReRAM.

Towards Efficient Neural Decoder for Dexterous Finger Force Predictions.

Dexterous control of robot hands requires a robust neural-machine interface capable of accurately decoding multiple finger movements. Existing studies primarily focus on single-finger movement or rely heavily on multi-finger data for decoder training, which requires large datasets and high computation demand. In this study, we investigated the feasibility of using limited single-finger surface electromyogram (sEMG) data to train a neural decoder capable of predicting the forces of unseen multi-finger combinations. We developed a deep forest-based neural decoder to concurrently predict the extension and flexion forces of three fingers (index, middle, and ring-pinky). We trained the model using varying amounts of high-density EMG data in a limited condition (i.e., single-finger data). We showed that the deep forest decoder could achieve consistently commendable performance with 7.0% of force prediction errors and R2 value of 0.874, significantly surpassing the conventional EMG amplitude method and convolutional neural network approach. However, the deep forest decoder accuracy degraded when a smaller amount of data was used for training and when the testing data became noisy. The deep forest decoder shows accurate performance in multi-finger force prediction tasks. The efficiency aspect of the deep forest lies in the short training time and small volume of training data, which are two critical factors in current neural decoding applications. This study offers insights into efficient and accurate neural decoder training for advanced robotic hand control, which has the potential for real-life applications during human-machine interactions.

The technology of precise tapping of remaining oil during the high water cut stage of offshore multi-layer reservoirs driven by big data intelligence

Abstract Based on traditional numerical simulation and optimization algorithm, combined with the “hard data” monitored by stratified injection and production real-time monitoring and automatic control technology, a fine reservoir model integrated reservoir knowledge and development law was established, and the remaining oil potential exploitation during high water cut stage was quantified and evaluated by deep forest method to adjust well productivity, greatly improving the remaining oil potential exploitation effect. The innovation has formed a set of accurate remaining oil potential tapping technology based on big data intelligent drive in offshore multi-layer reservoirs with high water cuts. This technology fully adopts the automatic and intelligent characteristics of data-driven and deep forest algorithms, and the historical fitting rate of production dynamic during the high water cut stage is more than 90%, which can accurately guide the design and implementation of the remaining oil potential tapping scheme and provide a reference for similar oil fields.

CVAE-DF: A hybrid deep learning framework for fertilization status detection of pre-incubation duck eggs based on VIS/NIR spectroscopy

Unfertilized duck eggs not removed prior to incubation will deteriorate quickly, posing a risk of contaminating the normally fertilized duck eggs. Thus, detecting the fertilization status of breeding duck eggs as early as possible is a meaningful and challenging task. Most existing work usually focus on the characteristics of chicken eggs during mid-term hatching. However, little attention has been paid to the detection for duck eggs prior to incubation. In this paper, we present a novel hybrid deep learning detection framework for the fertilization status of pre-incubation duck eggs, termed CVAE-DF, based on visible/near-infrared (VIS/NIR) transmittance spectroscopy. The framework comprises the encoder of a convolutional variational autoencoder (CVAE) and an improved deep forest (DF) model. More specifically, we first collected transmittance spectral data (400–1000 nm) of 255 duck eggs before hatching. The multiplicative scatter correction (MSC) method was then used to eliminate noise and extraneous information of the raw spectral data. Two efficient data augmentation methods were adopted to provide sufficient data. After that, CVAE was applied to extract representative features and reduce the feature dimension for the detection task. Finally, an improved DF model was employed to build the classification model on the enhanced feature set. The CVAE-DF model achieved an overall accuracy of 95.94 % on the test dataset. These experimental results in terms of four metrics demonstrate that our CVAE-DF method outperforms the traditional methods by a significant margin. Furthermore, the results also indicate that CVAE holds great promise as a novel feature extraction method for the VIS/NIR spectral analysis of other agricultural products. It is extremely beneficial to practical engineering.