A structural damage detection method by integrating data fusion and random forests was proposed. The original acceleration signals were translated into energy features wavelet packet decomposition. Then the processed fused new fusion. This can further enlarge differences among all types of damages. Finally, as an effective classifier used to detect multiclass damage. Numerical study on benchmark model eight-storey steel shear frame structure carried out validate accuracy proposed method....

To address these issues, this study proposes a lightweight and highly deployable vehicle detection model. The model builds upon the single-stage YOLOv7 architecture, replacing original convolutional layers with inverted residual structure depth-wise separable convolutions from MobileNetv3. This streamlines network width computational parameters. Furthermore, an ECA attention mechanism is incorporated into both backbone multi-scale feature branches to reduce overhead while enhancing model's...

Pathology plays a critical role in diagnosing wide range of diseases, yet existing approaches often rely heavily on task-specific models trained extensive, well-labeled datasets. These methods face sustainability challenges due to the diversity pathologies and labor-intensive nature data collection. To address these limitations, we highlight need for universal multimodal embeddings that can support multiple downstream tasks. Previous involve fine-tuning CLIP-based models, which handle images...

The rapid development of pay-as-you-go cloud services motivates the increasing number resource demands. However, volatile demands bring new challenges for current techniques to minimize cost capacity planning and VM provisioning while satisfying customer service vendors will incur enormous revenue loss within long-term inappropriate planning, especially when fluctuate abruptly frequently. In this paper, we cast as a classification problem propose an integrated framework, which effectively...

The gene expression data are usually provided with a large number of genes and relatively small samples, which brings lot new challenges. Selecting those informative becomes the main issue in microarray analysis. Recursive cluster elimination based on support vector machine (SVM-RCE) has shown better classification accuracy some sets than recursive feature (SVM-RFE). However, SVM-RCE is extremely time-consuming. In this paper, we propose an improved method called ISVM-RCE. ISVM-RCE first...

Vehicular nodes are equipped with more and sensing units, a large amount of data is generated. Recently, research considers cooperative urban as the heart intelligent green city traffic management. The key components platform will be combination pervasive vehicular system, well central control analysis where data-gathering fundamental component. However, monitoring also challenging issues in sensor networks because dynamic nature network. In this paper, we propose an efficient continuous...

Deep learning has been applied to structural damage detection and achieved great success in recent years, such as the popular methods based on vibration response convolutional neural networks (CNN). However, due limited number of samples that can be acquired practice for detection, CNN-based models may not fully trained; thus, their performance identifying different severity well locations reduced. To solve this issue, paper, we follow strategy "divide-and-conquer" propose a two-stage...

Gridding is the first and most important step to separate spots into distinct areas in microarray image analysis. Human intervention necessary for gridding methods, even if some so-called fully automatic approaches also need preset parameters. The applicability of these methods limited certain domains will cause variations gene expression results. In addition, improper gridding, which influenced by both misalignment high noise level, affect throughput this paper, we have presented a...

Load prediction is a crucial task in the area of cloud computing resource management. Due to ever-changing characteristics workload data Kubernetes clusters, influenced by time and varying access patterns different applications, these uncertainties present significant challenges series data. With rapid advancement artificial intelligence technology, deep learning methods based on Transformer architecture have begun gradually replace traditional statistical approaches machine algorithms for...

Semi-supervised cluster ensemble usually introduces a small amount of supervision in the first stage ensemble, i.e., generation, by performing many runs semi-supervised clustering algorithms. However, it is neither efficient terms computational complexity, nor flexible dynamic learning environment where limited changes over time. In this article we propose new framework which generates base partitions an unsupervised manner and attributes different weights to each partitions. The weighting...

Selection of relevant genes for sample classification is a common task in most gene expression studies. As powerful approach, random forest has been applied this field, and it shows excellent performance compared with other methods. The measure variable importance the key selection using forest. However, existing methods just consider original based on OOB error. In paper, we proposed new difference proximity matrix, used DNA microarray data. Compared analysis forest, method more sensitive...

Abstract. City hotspots refer to the areas where residents visit frequently, and large traffic flow exist, which reflect people travel patterns distribution of urban function area. Taxi trajectory data contain abundant information about functions citizen activities, extracting interesting city from them can be importance in planning, command, public services etc. To detect discover a variety changing among them, we introduce field-based cluster analysis technique pick-up drop-off points taxi...

Abstract. A method of trajectory clustering based on decision graph and data field is proposed in this paper. The utilizes to describe spatial distribution points, uses discover cluster centres. It can automatically determine parameters suitable clustering. applied taxi data, which are the holiday (May 1st, 2014), weekday (Wednesday, May 7th, 2014) weekend (Saturday, 10th, respectively, Wuhan City, China. hotspots four hours (8:00-9:00, 12:00-13:00, 18:00-19:00 23:00-24:00) for three days...

With the development of power monitoring technology, more and real-time data are accumulated in system. Due to large amount high dimension data, traditional analysis methods often unable discover hidden rules complex relationships between fault reasons. In recent years, mining techniques have been successfully applied many fields. Compared with statistical commonly used expert system, data-driven based also show excellent performance detection systems. However, most these can only detect...