Abstract With the rapid growth of mobile internet and smart devices, China's mobile game market has become highly competitive. Game companies face increasing costs in promoting games and acquiring new customers. In response, this paper proposes a stack-based model for game customer churn prediction. This model utilizes a comprehensive set of user features, including static, dynamic, and social network data extracted from game logs. To correct the imbalance between positive and negative examples in our dataset, we applied the Neighborhood Cleaning Rule (NCL). We then constructed the model using a stacking approach that integrates multiple machine learning models into a cohesive binary classification framework. This method systematically predicts customer churn by processing these features. Our approach was rigorously tested on a large-scale real-world dataset, comparing it against nine standard baseline methods. The results reveal that the stack-based model significantly outperforms these methods, achieving a prediction accuracy of 94%.

Aiming at the problem that the stiffness performance of industrial robots depends on the posture, a method to improve the stiffness performance by optimizing the robot posture is proposed based on the robot posture stiffness evaluation index. The robot joint stiffness identification model is established by using the principle of virtual work and Jacobi matrix, and the stiffness values of six joints are obtained by least-squares. On the basis of revealing the mapping relationship between robot joints and end, the robot end flexibility ellipsoid is obtained, and the volume of the end flexibility ellipsoid is utilized as the global stiffness coefficient of the robot, so as to obtain the robot posture stiffness evaluation index. According to the posture stiffness evaluation index, the robot posture optimization model is established, and the optimized posture is obtained by using genetic algorithm. Taking the QJR6-1 robot as the experimental object, under the maximum load, the deformation of the robot end center is reduced from 0.6238 to 0.3984 mm, which is 36.13%, and the deformation of the end is greatly reduced compared with that before the optimization, which indicates that the robot’s stiffness is improved after the posture optimization, and it verifies the feasibility of the optimization model, and provides a reference to improve the stiffness performance of industrial robots further.

Introduction: Body measurement traits are integral in cattle production, serving as pivotal criteria for breeding selection. Wenshan cattle, a local breed in China’s Yunnan province, exhibit remarkable genetic diversity. However, the molecular mechanisms regulating body measurement traits in Wenshan cattle remain unexplored.Methods: In this study, we performed a genome-wide association method to identify genetic architecture for body height body length hip height back height (BAH), waist height and ischial tuberosity height using the Bovine 50 K single nucleotide polymorphism Array in 1060 Wenshan cattles.Results: This analysis reveals 8 significant SNPs identified through the mixed linear model (MLM), with 6 SNPs are associated with multiple traits and 4 SNPs are associated with all 6 traits. Furthermore, we pinpoint 21 candidate genes located in proximity to or within these significant SNPs. Among them, Scarb1, acetoacetyl-CoA synthetase and HIVEP3 were implicated in bone formation and rarely encountered in livestock body measurement traits, emerge as potential candidate genes regulating body measurement traits in Wenshan cattle.Discussion: This investigation provides valuable insights into the genetic mechanisms underpinning body measurement traits in this unique cattle breed, paving the way for further research in this domain.

With the development of hyperspectral imaging technology, the potential for utilizing hyperspectral images to accurately estimate heavy metal concentrations in regional soil has emerged. Currently, soil heavy metal inversion based on laboratory hyperspectral data has demonstrated a commendable level of accuracy. However, satellite images are susceptible to environmental factors such as atmospheric and soil background, presenting a significant challenge in the accurate estimation of soil heavy metal concentrations. In this study, typical chromium (Cr)-contaminated agricultural land in Shaoguan City, Guangdong Province, China, was taken as the study area. Soil sample collection, Cr content determination, laboratory spectral measurements, and hyperspectral satellite image collection were carried out simultaneously. The Zhuhai-1 hyperspectral satellite image spectra were corrected to match laboratory spectra using the direct standardization (DS) algorithm. Then, the corrected spectra were integrated into an optimal model based on laboratory spectral data and sample Cr content data for regional inversion of soil heavy metal Cr content in agricultural land. The results indicated that the combination of standard normal variate (SNV)+ uninformative variable elimination (UVE)+ support vector regression (SVR) model performed best with laboratory spectral data, achieving a high accuracy with an R2 of 0.97, RMSE of 5.87, MAE of 4.72, and RPD of 4.04. The DS algorithm effectively transformed satellite hyperspectral image data into spectra resembling laboratory measurements, mitigating the impact of environmental factors. Therefore, it can be applied for regional inversion of soil heavy metal content. Overall, the study area exhibited a low-risk level of Cr content in the soil, with the majority of Cr content values falling within the range of 36.21-76.23 mg/kg. Higher concentrations were primarily observed in the southeastern part of the study area. This study can provide useful exploration for the promotion and application of Zhuhai-1 image data in the regional inversion of soil heavy metals.

This research aimed to design and implement a wireless monitoring and data acquisition system using LabVIEW for a 3 kW Solar Fuel Cell Reactor. The system incorporated advanced technologies like Raspberry Pi 3B+, Arduino Nano, PT100 temperature sensors, voltage divider, ACS758, and calibrated cells for precise irradiance sensing. It adhered to the IEC-61724-2017 standard and utilized a 3D-printed housing with a DIN rail fastener for durability and convenience. During a 5-day test period, the study examined the impact of temperature on the photovoltaic module, revealing variations of up to 3 °C among cells, leading to parameter dispersion and performance losses. Temperature also showed an inverse relationship with voltage, with higher temperatures resulting in decreased efficiency and a maximum surface temperature of 52.31 °C. The research also explored the effect of irradiance, finding a direct correlation between irradiance and generated current. Notably, extreme solar irradiance events occurred, with the highest value reaching 1245.90 W/m2 for 6 s at 11:39:13 on June 10, 2023, and the longest event lasting 176 s at 1219.77 W/m2, occurring at 11:34:17. The collected indicators provided reliable information on energy generation by the SFCR under specific environmental conditions.

Background Ultrasound image segmentation is challenging due to the low signal-to-noise ratio and poor quality of ultrasound images. With deep learning advancements, convolutional neural networks (CNNs) have been widely used for ultrasound image segmentation. However, due to the intrinsic locality of convolutional operations and the varying shapes of segmentation objects, segmentation methods based on CNNs still face challenges with accuracy and generalization. In addition, Transformer is a network architecture with self-attention mechanisms that performs well in the field of computer vision. Based on the characteristics of Transformer and CNNs, we propose a hybrid architecture based on Transformer and U-Net with joint loss for ultrasound image segmentation, referred to as TU-Net. Methods TU-Net is based on the encoder-decoder architecture and includes encoder, parallel attention mechanism and decoder modules. The encoder module is responsible for reducing dimensions and capturing different levels of feature information from ultrasound images; the parallel attention mechanism is responsible for capturing global and multiscale local feature information; and the decoder module is responsible for gradually recovering dimensions and delineating the boundaries of the segmentation target. Additionally, we adopt joint loss to optimize learning and improve segmentation accuracy. We use experiments on datasets of two types of ultrasound images to verify the proposed architecture. We use the Dice scores, precision, recall, Hausdorff distance (HD) and average symmetric surface distance (ASD) as evaluation metrics for segmentation performance. Results For the brachia plexus and fetal head ultrasound image datasets, TU-Net achieves mean Dice scores of 79.59% and 97.94%; precisions of 81.25% and 98.18%; recalls of 80.19% and 97.72%; HDs (mm) of 12.44 and 6.93; and ASDs (mm) of 4.29 and 2.97, respectively. Compared with those of the other six segmentation algorithms, the mean values of TU-Net increased by approximately 3.41%, 2.62%, 3.74%, 36.40% and 31.96% for the Dice score, precision, recall, HD and ASD, respectively.