Aiming at the problems of poor point cloud data fusion in traditional MLP models, this paper proposes a multimodal 3D target detection network based on KANs. A KANDyVFE encoder incorporating a fusion layer is designed with KANs as the backbone, and a self-attention mechanism is used to dynamically fuse point cloud features. Two datasets, KITTI and WaymoOpen, are selected as 3D target detection datasets to explore the performance level of the algorithm through controlled experiments. Based on ablation experiments, the effectiveness of the KANDyVFE encoder and the self-attention fusion module is verified. The proposed algorithm achieves 80.72% and 80.23% 3DmAP and 3DmAPH on the WaymoOpen dataset for LEVEL_1, which is 2.14% and 2.17% better than the closest BtcDet method, and achieves the same advanced performance on LEVEL_2. When the KANDyVFE encoder module is not used, the 3DmAP and 3DmAPH are only 72.36% and 74.35%, respectively, and the addition of the KANDyVFE encoder and the self-attention fusion module achieves 91.33% and 92.09% for 3DmAP and 3DmAPH, respectively. The experimental results validate the effectiveness of KANs in point cloud applications, and the ablation experiments further demonstrate the performance improvement brought by the designed modules.

Thanks to the wave of digital economic globalization, the business development of cross-border e-commerce platforms is in full swing. This paper aims to promote the development of e-commerce personalization and launch the research of consumer behavior characteristics. This paper utilizes the concept of entropy in information theory to modify the weights of user feature vectors, so as to make up for the inadequacy of the K-Means algorithm in expressing ambiguous clustering information. Combined with the data samples, the consumer behavior prediction model is established. For the dynamic clustering of customer groups, construct the customer segmentation model based on the improved K-Means algorithm. Combined with the time series prediction model, complete the formation of the spatio-temporal data mining model of consumer behavior. The model is used to mine the consumer behavior dataset of a cross-border e-commerce platform, and the clustering analysis yields four precise consumer group portraits. In this paper, by mining and analyzing the characteristics of consumer spatio-temporal data, the cross-border e-commerce platform is provided with more accurate user insights and marketing optimization solutions.

This paper takes ten economies as examples to analyze and assess the current situation of their international trade development through RCA, MS and TC indexes. On the basis of Porter’s “diamond model” theory, a comprehensive evaluation index system of international trade competitiveness is set up in combination with the actual situation. The entropy value method is used to measure the comprehensive index of international trade competitiveness, and the influence of various influencing factors on international trade competitiveness is empirically studied based on the principal component multiple regression analysis. The results show that the U.S. international trade competitiveness is far ahead, with an average score of 3.67 in 2020-2024, and the lowest score is Singapore, with a score of only -2.17. The degree of explanation of international trade competitiveness of the four factors reaches 98.9%, and all of them have a promotional effect on the international trade competitiveness, in the following order: factors of production>enterprise strategy and competition>related industries>demand factors.

The construction and opening of high-speed railroads have brought new development opportunities to China’s ethnic regions, which are economically backward but rich in tourism resources. From the perspective of the impact of high-speed rail on regional tourism, this paper briefly analyzes the homogenization effect and accessibility effect that the construction of high-speed rail brings to the corresponding region. Accordingly, it puts forward the relevant research hypotheses on the impact of high-speed rail on regional tourism and analyzes the current situation of tourism market development in China’s A ethnic region. Under this premise, the model of high-speed railroad influence on tourism development level is designed and relevant research variables are selected. Based on the model, the empirical analysis of the impact of high-speed railroad on tourism in ethnic region A is launched. The study points out that the opening of high-speed railroad significantly promotes the total tourism income of ethnic region A at the 1% level, i.e., the opening of high-speed railroad has a positive positive effect on the tourism development of ethnic regions.

Aiming at the complexity of mental health assessment for students in colleges and universities, this paper proposes an innovative framework that integrates social sentiment analysis and multi-branch neural networks. A multilevel mental health assessment system is constructed through cross-modal feature interaction CNN+BiGRU with heterogeneous graph structure modeling. In the model design, image feature extraction is pre-trained by five-branch CNN structure ViT, text features are fused by dynamic word embedding with multi-scale convolution, and a virtual node and metapath-driven heterogeneous graph neural network H-GNN is introduced to strengthen the global relationship modeling. Experiments show that the model achieves 89.7% and 91.2% accuracy on Twitter-15 and Twitter-17 datasets, respectively, and the F1 values are improved by 3.24% and 2.32% from the optimal baseline BICCM. In the actual college mental health monitoring, the model successfully captured the time-series fluctuations of depression index and anxiety level, and found that the rational-perceptual dimension was highly correlated with the examination cycle, with 0.69 during the midterm examination and 0.68 during the final examination. Through the ten-fold cross-validation comparison experiments, the model significantly outperforms the cutting-edge models, such as MIMNBERT, EF-NET and so on on the weighted average index, with an average accuracy rate of 99.02% and F1 value of 98.08%. The study shows that the framework provides a highly accurate and interpretable technical solution for mental health risk early warning, which is especially suitable for dynamic monitoring scenarios in universities.

Taking the perspective of new quality productivity, this study explores the promotion effect of the intermingling of intelligent computing and traditional culture on the cultivation of innovative talents, and constructs an evaluation system containing four primary indicators and 14 secondary indicators of educational activities, student practice, collaborative innovation and teaching resources. The cloud integration model is used to deal with the ambiguity and randomness of the complex system, and the network hierarchy analysis method ANP is used to determine the weights of the indicators and reveal the dynamic association of each element. It is found that: the indicator B2 of student practice category has the highest weight of 0.329, in which the number of awards of C5 innovation competition and the number of C4 students’ project participation are the core driving factors, with the weights of 0.103 and 0.078, respectively. the cloud integration model verifies the scientificity of the evaluation system. The evaluation value of the traditional culture innovation talent evaluation system constructed in this paper is 0.798, and the integrated cloud model belongs to “very good” grade. However, the mapping intervals of C14 Resource Library Call Frequency and C13 Teacher Integration Background are low, 0.346 and 0.413 respectively, which need to be adjusted and optimized. The innovative talent cultivation program of colleges and universities constructed in this study can make up for the shortcomings in traditional talent cultivation performance evaluation, has certain practicality and effectiveness, and helps to improve the quality of traditional culture innovative talent cultivation.

Under the background of economic transformation and high-quality development strategy, the coordinated development of regional economy and precise regulation of fiscal policy have become a hot spot of concern. This paper constructs the evaluation index system of regional economic high quality development and evaluates the level of economic high-quality development in the recent 10 years by using the entropy weight-TOPSIS method. Combined with Dagum Gini coefficient and spatial autocorrelation test, we study the spatial correlation of economic high-quality development among regions. The benchmark regression model and mediation effect model are constructed to calculate the effect of fiscal policy on the level of regional economic high-quality development, and to judge the effectiveness of fiscal policy regulation path. The study shows that the 30 provinces in China can be divided into different regions according to the level of high-quality economic development, and each region presents different types of characteristics such as “high – high”, with significant differences. The variables in the benchmark regression model and the mediation effect model are correlated at the 1% level, and pass the smoothness test when the difference is of the 0th order. Fiscal policy at the regional level has a positive contribution to the level of high-quality economic development, but at the same time is affected by the original level of development of each region.

The article takes the defect detection and recognition of railroad track as the main research point, and extracts, preprocesses and corrects the railroad track surface image by introducing image segmentation algorithm. Gabor function, K-means clustering method and conditional iterative pattern algorithm are embedded in the original Markov random field model to construct the improved two layer graph model for railroad track defect segmentation. The recall, precision, mean average precision, and loss function of the improved Markov defect segmentation model are significantly better than those of the original model, and the mean average precision of the defect segmentation model is increased to 95.7% after the Gabor function, K-means clustering method, and conditional iterative pattern algorithm are applied. The improved Markov defect segmentation model fused with clustering features in this paper can better meet the classification and identification of railroad track defects.

In recent years, with the rapid development of artificial intelligence, big data, machine learning and other technologies, human society is entering a more and more intelligent society, and the interaction between humans and machines becomes more and more common. In this paper, image processing operations are added on the basis of Kinect’s original acquisition of gong dance images, which reduces the influence of external light, background and other factors, and makes the human capture efficiency increase dramatically, and a spatio-temporal graph is constructed on the basis of the continuous human posture key point data, which describes the distribution of the human posture key points in different dataset types. Aiming at the problems existing in the traditional spatio-temporal map convolutional network, a multi-dimensional attention mechanism is designed to guide the model to reasonably allocate the weight resources in three dimensions: space, time and channel, respectively. Experiments are conducted on NTU-RGB+D, Kinect skeleton and Taiji datasets, respectively, which show that the AGCN-STC proposed in this paper has better recognition performance on all three datasets, and the recognition accuracy is improved by 0.9 percentage points compared with AM-GCN. Two actors are used as samples for visual measurement and quantitative analysis to compare the differences between the performance gestures of the two ornaments. Finally, based on the results of the study, we propose a transmission path for the Guanzhong gong dance, which is a reference for the cultural transmission of the Guanzhong gong dance.

In the process of social development of Tang Dynasty, literary works behind the depth of interpretation and expression, systematized spiritual concepts. In this paper, the text data of Tang Dynasty literary works are processed by word division and de-discontinued words, and it is intended to use Transformer model to realize the word vector transformation of text data, and put the word vector into Text-CNN network for iterative training to realize the text feature extraction. By means of text feature screening, the cultural value assessment system of Tang Dynasty literary works is formed, and a comprehensive evaluation model of cultural value is designed under the role of convolutional neural network and text features, and using the model of this paper, the cultural value of Tang Dynasty literary works is assessed. The accuracy rate of cultural value classes “Ⅱ”, “Ⅲ” and “Ⅴ” is 1, while the accuracy rate of cultural value classes “I” and “Ⅳ” have accuracy rates of 0.98 and 0.96, indicating that the model in this paper can accurately assess the cultural value in Tang Dynasty literary works.