Dynamic pick-up point recommendation with multi-modal deep forest and incentive-based adaptive Kuhn-Munkres Algorithm

E-commerce platforms such as Taobao can collect massive users 'shopping behavior data, which makes it possible to grasp users' shopping preferences. However, the current research methods of user operation behavior prediction usually only analyze a certain type of user's operation behavior, which cannot fully reflect the overall characteristics of user behavior. Based on the shopping behavior data of Alibaba's e-commerce platform, this article mines user characteristics, product characteristics, product category characteristics, user-product characteristics, and user-product category characteristics from a large amount of online shopping behavior data. The online purchasing behavior prediction model has achieved better results than other models. Experimental results show that the model improves the accuracy of prediction while reducing the time overhead.

Dioxins (DXN) is a persistent environmental pollutant that poses risks such as a weakened immune system, and teratogenic and carcinogenic effects. Municipal solid waste incineration (MSWI) plants are one of the major DXN generation sources. It is imperative to implement the monitoring and control. However, the harsh environment prevents the use of conventional equipment for detection, resulting in a lack of information on DXN generation concentration. This article presents an advanced tree-based interpretable deep modeling approach that utilizes a multimodal data-driven strategy. The available data types include two modalities: numerical and image data. To address the above issue and modeling, first, the time scale of the multimodal data is adjusted to match the sampling period of DXN based on the mechanism knowledge. Then, a novel adaptive deep forest regression algorithm based on cross-layer full connection (ADFR-clfc) is proposed for modeling process numerical data and recorded operational data. Furthermore, a convolutional neural network feature extraction method based on transfer learning combined with ADFR-clfc is employed for modeling image data. Finally, the DXN generation concentration is obtained by taking the arithmetic average of the former models. The proposed method is validated using approximately one year of data in an MSWI plant in Beijing. Experimental results show that the root mean square error (RMSE) of the concentration estimate is 0.0864 and the MAE is 0.0707, demonstrating the effectiveness of the proposed method.

In order to realize UAV fault detection and recovery in complex flight modes, a data-driven FCM-(PSO-DF) fusion algorithm is proposed which consists of fuzzy c-means (FCM), particle swarm optimization (PSO) and deep forest (DF) algorithm. Firstly, the flight data is preprocessed to remove abnormal values and high-frequency noise interference. Then, FCM clustering algorithm is used to divide the flight data into four clusters of samples according to different flight modes. Secondly, PSO-DF is designed as the regression model, and four groups of clustering samples are trained respectively to obtain the estimation of monitoring parameters. Then, the residual error between the actual flight data and the estimated value is calculated, the mean value of residuals is set as statistical threshold, and the residual error is compared with the statistical threshold to realize fault detection. Finally, the fault data is replaced with estimates to achieve fault recovery. In order to verify the effectiveness of this method, experiments are carried out using real flight data, and set up the comparison experiments with other algorithms. Experimental results show that the algorithm is effective on the failure detection and recovery of UAV flight data.

This research addresses the need for a predictive approach to detect security breaches in AI-powered mobile cloud applications. We propose a novel approach combining Radial ResNet for advanced feature extraction with Random Forest (RF) for classification. This hybrid model is designed to analyze complex and dynamic datasets in real-time, enhancing the predictive capabilities for identifying potential security threats. Results indicate significant efficacy, with the algorithm achieving high accuracy and sensitivity in predicting security breaches. The Radial ResNet–RF combination achieved a training accuracy of 98.5%, with precision, recall, and F1-score values of 97.8%, 98.2%, and 98.0%, respectively. On testing and validation datasets, the model demonstrated accuracies of 96.2% and 95.5%, respectively.

This research addresses the need for a predictive approach to detect security breaches in AIpowered mobile cloud applications. We propose a novel approach combining Radial ResNet for advanced feature extraction with Random Forest (RF) for classification. This hybrid model is designed to analyze complex and dynamic datasets in real-time, enhancing the predictive capabilities for identifying potential security threats. Results indicate significant efficacy, with the algorithm achieving high accuracy and sensitivity in predicting security breaches. The Radial ResNet-RF combination achieved a training accuracy of 98.5%, with precision, recall, and F1score values of 97.8%, 98.2%, and 98.0%, respectively. On testing and validation datasets, the model demonstrated accuracies of 96.2% and 95.5%, respectively.

A new adaptive weighted deep forest algorithm which can be viewed as a modification of the confidence screening mechanism is proposed. The main idea underlying the algorithm is based on adaptive weigting of every training instance at each cascade level of the deep forest. The confidence screening mechanism for the deep forest proposed by Pang et al., strictly removes instances from training and testing processes to simplify the whole algorithm in accordance with the obtained random forest class probability distributions. This strict removal may lead to a very small number of training instances at the next levels of the deep forest cascade. The presented modification is more flexible and assigns weights to instances in order to differentiate their use in building decision trees at every level of the deep forest cascade. It overcomes the main disadvantage of the confidence screening mechanism. The proposed modification is similar to the AdaBoost algorithm to some extent. Numerical experiments illustrate the outperformance of the proposed modification in comparison with the original deep forest. It is also illustrated how the proposed algorithm can be extended for solving the transfer learning and distance metric learning problems.

Visual simultaneous localization and mapping (SLAM) algorithms in dynamic scenes will apply the moving feature points to the camera pose’s calculation, which will cause the continuous accumulation of errors. As a target-detection tool, mask R-CNN, which is often used in combination with the former, due to the limited training datasets, easily results in the semantic mask being incomplete and deformed, which will increase the error. In order to solve the above problems, we propose in this paper a visual SLAM algorithm based on an adaptive mask dilation strategy and the dynamic probability of the feature points, named MDP-SLAM. Firstly, we use the mask R-CNN target-detection algorithm to obtain the initial mask of the dynamic target. On this basis, an adaptive mask-dilation algorithm is used to obtain a mask that can completely cover the dynamic target and part of the surrounding scene. Then, we use the K-means clustering algorithm to segment the depth image information in the mask coverage area into absolute dynamic regions and relative dynamic regions. Combined with the epipolar constraint and the semantic constraint, the dynamic probability of the feature points is calculated, and then, the highly dynamic possible feature points are removed to solve an accurate final pose of the camera. Finally, the method is tested on the TUM RGB-D dataset. The results show that the MDP-SLAM algorithm proposed in this paper can effectively improve the accuracy of attitude estimation and has high accuracy and robustness in dynamic indoor scenes.

Apart from being relied upon for feeding the entire world, the agricultural sector is also responsible for a third of the global Gross-Domestic-Product (GDP). Additionally, a majority of developing nations depend on their agricultural produce as it provides employment opportunities for a significant fraction of the poor. This calls for methods to ensure the accurate and efficient diagnosis of plant disease, to minimize any adverse effects on the produce. This paper proposes the recognition and classification of maize plant leaf diseases by application of the Deep Forest algorithm. The Automated novel approach and accurate classification using the Deep Forest technique are a significant step-up from the existing manual classification and other techniques with less accuracy. The proposed approach has outperformed Deep Neural models and other traditional machine learning algorithms in terms of accuracy. It justifies its low dependency on extensive Hyper-parameter tuning and the size of the dataset as against other Deep Learning Models based on neural networks.

Accurate prediction of wind speed is a prerequisite for the safe and accurate operation of wind power generation, however, WRF models typically do not produce sufficiently accurate wind speed predictions. This study proposed a Seasonal and Temporal Correlation - Deep Forest (STC-DF) model for offshore wind speed prediction. Different from traditional methods, the STC-DF model takes the advantages of the deep forest algorithm to automatically learn complex feature interactions without manual feature engineering. The model is designed to capture the seasonal and temporal characteristics of wind speed variations. To test the effectiveness of the proposed method, we applied the trained STC-DF model to an offshore wind farm in Hainan Province, China. Seven days of data from each season were selected for testing. The results show that the STC-DF model can effectively reduce the error caused by WRF forecast. The error index of the corrected wind speed reduced more than 40%, the accuracy of wind speed forecast increased 15%. And the method passed the multi-model comparison test and robustness experiment. These research results show that the STC-DF model has strong versatility and good correction ability, and is suitable for wind speed forecasting in different regions, which is a feasible method to improve the reliability of offshore wind power generation.

Electronics-rich analog systems are difficult to diagnose owing to their complex working mechanisms and the variability of the working environment. In recent years, deep learning has been gradually applied to the field of circuit system fault diagnosis because of its strong ability to mine the intrinsic characteristics of signals. However, the traditional deep learning method requires a lot of effort to achieve satisfactory results due to large number of parameters, complex models, slow training speed, and large datasets. The key factors for the success of traditional deep learning methods are layer-by-layer processing, feature transformation within the model, and sufficient model complexity. Deep forest (DF) is a new feature learning model that inherits the three characteristics of the traditional deep learning model but is that it is not based on neural network. It has fewer hyperparameters, a simpler model, faster training speed. In this article, an improved DF algorithm based on nonparametric predictive inference (NPI) is proposed, named NPIDF, which can better deal with small sample data. In two typical analog filter circuit fault diagnosis experiments, it is proved that DF and NPIDF achieve good diagnosis effect, and NPIDF performance is better, showing a greater advantage in small sample data.

Due to the richness of biological and drug data, identifying hidden DNA binding sites has been a rewarding but challenging task. To learn DNA binding site features efficiently, we propose a WeighDF algorithm based on combined feature encoding and weighted multi-granularity scans. The algorithm WeighDF assigns a unique weight vector to the scanning window during multi-granularity scanning, which emphasizes more important features in the feature sequence and reduces the interference from other features. Experimental results show that the algorithm WeighDF achieved higher accuracy on two different datasets, reaching 0.8943, which is an improvement of about 1.23% compared to deep forests. In general, the algorithm WeighDF outperforms the traditional deep forest algorithm across all evaluation indicators and demonstrates better portability and higher prediction accuracy. This research can not only detect hidden DNA binding sites but also provide insights for the study of efficient recombination systems in drug design.

A material handling scheduling method for mixed-model automotive assembly lines based on an improved static kitting strategy

<div class="section abstract"><div class="htmlview paragraph">The application of millimeter-wave radar technology in autonomous driving has become increasingly widespread with the rapid development of intelligent transportation systems. However, millimeter-wave radar is easily affected by environmental noise, multipath reflections, and electromagnetic interference, resulting in a large number of invalid target signals that reduce the system’s detection accuracy and safety. We proposes a method for filtering invalid targets based on interference signal characteristics and an Adaptive Interactive Multiple Model Kalman Filter (IMM-KF) target tracking algorithm. First, we effectively filter out empty targets, ghost targets, and false targets through a threshold method and lifecycle assessment, achieving a filtering rate exceeding 99.8%. Second, the improved Adaptive IMM-KF algorithm, combined with the Hungarian algorithm, associates and tracks multiple targets. The root mean square error (RMSE) of our methods is reduced by 7.07% and 8.05% compared to the traditional IMM and Unscented Kalman Filter (UKF) algorithms in scenarios with a single pedestrian moving in a straight line. And in scenarios involving both pedestrians and motor vehicles, the RMSE is reduced by 4.29% and 7.61% compared to the traditional IMM and UKF algorithms, respectively. Real-vehicle experiments have validated the robustness and accuracy of the proposed method across various scenarios.</div></div>

In dynamic environments, moving objects introduce unstable features that significantly degrade the accuracy of simultaneous localization and mapping (SLAM) systems. To address this issue, we propose Neural-KF, a robust visual SLAM framework that integrates three key modules: (1) a modified SuperPoint network with multi-level feature fusion for reliable static keypoint extraction, (2) a YOLOv8-based dynamic object detector, and (3) a Kalman-consistent state estimation mechanism that predicts object motion trajectories to enhance temporal consistency. By associating predicted and detected bounding boxes via the Hungarian algorithm, Neural-KF achieves accurate suppression of dynamic points while preserving sufficient static features for pose estimation. Experimental evaluations on public datasets, including KITTI and EuRoC, demonstrate that Neural-KF improves absolute trajectory error by up to 28% compared to VINS-Fusion and achieves competitive accuracy against advanced dynamic SLAM systems such as DynaSLAM. Furthermore, the system maintains real-time performance (>30 FPS) with a balanced trade-off between accuracy and computational cost. These results highlight the effectiveness of Neural-KF in achieving robust and efficient visual odometry under challenging dynamic conditions.

Volumetric media has the potential to provide the six degrees of freedom (6DoF) required by truly immersive media. However, achieving 6DoF requires ultra-high bandwidth transmissions, which real-world wide area networks cannot provide today. Therefore, recent efforts have started to target efficient delivery of volumetric media, using a combination of compression and adaptive streaming techniques. It remains, however, unclear how the effects of such techniques on the user perceived quality can be accurately evaluated. In this paper, we present the results of an extensive objective and subjective quality of experience (QoE) evaluation of volumetric 6DoF streaming. We use PCC-DASH, a standards-compliant means for HTTP adaptive streaming of scenes comprising multiple dynamic point cloud objects. By means of a thorough analysis, we investigate the perceived quality impact of the available bandwidth, rate adaptation algorithm, viewport prediction strategy and user's motion within the scene. We determine which of these aspects has more impact on the user's QoE, and to what extent subjective and objective assessments are aligned.