Composing fashion outfits involves deep understanding of standards while incorporating creativity for choosing multiple items (e.g., Jewelry, Bag, Pants, Dress). In websites, popular or high-quality are usually designed by experts and followed large audiences. this paper, we propose a machine learning system to compose automatically. The core the proposed automatic composition is score outfit candidates based on appearances meta-data. We leverage popularity oriented websites supervise...

Among the most popular micro-blogging service, Twitter recently introduced their reblogging service called retweet to allow a user repopulate another user's content for his followers. It quickly becomes one of prominent features on and an important mean secondary promotion. However, it remains unclear what motivates users whether retweeting decisions are predictable based tweeting history social relationships. In this paper, we propose modeling patterns using conditional random fields with...

Benchmark contamination has become a significant concern in the LLM evaluation community. Previous Agents-as-an-Evaluator address this issue by involving agents generation of questions. Despite their success, biases methods remain largely unexplored. In paper, we present theoretical formulation bias, providing valuable insights into designing unbiased protocols. Furthermore, identify two type bias through carefully designed probing tasks on minimal setup. To these issues, propose Unbiased...

As AI becomes increasingly embedded in the sharing economy, understanding its impact on consumer behavior is crucial. This research aims to examine how different human–AI relationship types—equal vs. human-dominant—influence responsible consumption, framed within Social Identity Theory. Specifically, we investigate mediating role of social identification and moderating effect anthropomorphism shaping responses interactions. Across three experimental studies, demonstrate that (1) equal...

Twitter has increasingly become an important source of information during disasters. Authorities have responded by providing related in Twitter. The same channel can also be used to deliver disaster preparation increase the readiness general public. Retweeting is key mechanism facilitate this diffusion process. Understanding factors that affect twitter users' retweet decision would help authority adopt optimal strategy choosing content, style, words, initial targeted users, time and...

Medical named entity recognition (NER) is an area in which medical entities are recognized from texts, such as diseases, drugs, surgery reports, anatomical parts, and examination documents. Conventional NER methods do not make full use of un-labelled texts embedded To address this issue, we proposed a approach based on pre-trained language models domain dictionary. First, constructed dictionary by extracting labelled collecting other resources, the Yidu-N4K data set. Second, employed to...

In social systems, agents often have different ability to persuade neighbors adopt their opinions. this paper, we aim investigate how the location and heterogeneity of influencers in networks can improve convergence. We propose a voter model with dynamic self-conviction heterogeneous individual influence which is related underlying network topology. An agent may keep its current opinion according personal conviction, or otherwise, it preferentially choose neighbor that has great influence....

Compressor fault diagnosis requires expert knowledge. Using the sequence labeling technology, this knowledge can be automatically extracted from compressor maintenance log sheets. Previous studies indicate that methods often need a substantial amount of annotation data for extraction, Unfortunately, are very scarce in field diagnosis. In paper, we introduce benchmark dataset extraction suitable air First, collected 11,418 pieces information Fault description, service requests, causes and...

Purpose This study leverages theories of social influence to explore how “likes” for consumption-related content on media fulfill consumers’ needs acceptance, subsequently affecting their repurchase and word-of-mouth (WOM) intentions. It aims understand the extent which engagements, specifically likes, serve as markers validation in context consumer behavior. Design/methodology/approach Our mixed-methods approach incorporated two experiments an analysis archival dataset from a popular...

Community detection is a pivotal task in data mining, and users' emotional behaviors have an important impact on today's society. So it very significant for society management or marketing strategies to detect communities social networks. Based the homophily of users networks, could confirm that would like gather together form according similarity. This paper exploits multivariate measure similarity, then takes advantage similarity as edge weight remodel network communities. The detailed...

The classification and visualization for surface objects receives a great deal of attention high spectral dimensional data processing. A lot methods were proposed applied in this problem over the past decade. Whereas most them still exist some challenge issues, include pre-treatment fussily, features extraction simplify, larger processing difficultly inaccurately. To solve these problems, we propose novel method bases on deep learning which combines maximum noise fraction (MNF) multilayers...

State-of-the-art image denoisers exploit various types of deep neural networks via deterministic training. Alternatively, very recent works utilize reinforcement learning for restoring images with diverse or unknown corruptions. Though can generate effective policy operator selection architecture search in restoration, how it is connected to the classic training solving inverse problems remains unclear. In this work, we propose a novel denoising scheme Residual Recovery using Reinforcement...

Feature extraction is one of the most important and effective method to reduce dimension in data mining, with emerging high dimensional such as microarray gene expression data. for selection, mainly serves two purposes. One identify certain disease-related genes. The other find a compact set discriminative genes build pattern classifier reduced complexity improved generalization capabilities. Depending on purpose types feature algorithms including ranking-based set-based are employed...

The outlier detection in deformation is always a hard problem to solve. As the requirements on automation and accuracy becoming stronger stronger, it also more important detect remove outliers monitoring observations as fast possible. In paper approximation of nonlinear function mapping relation using Artificial Neural Network (ANN) was introduced, issues about BP NN were discussed. To overcome drawbacks NN, Genetic Algorithm (GA) introduced into method reduce shortcomings much Aiming at...

The stock market refers to a financial in which individuals and institutions engage the buying selling of shares publicly listed firms. valuation stocks is influenced by interplay between forces supply demand. act allocating funds entails certain degree risk, while it presents possibility substantial gains over an extended period. task predicting prices securities further complicated presence non-stationary non-linear characteristics time series data. While traditional techniques have...

Deep depletion electric vehicle battery packs have a very large voltage range. The ultra wide range makes the optimal design of LLC converter extremely difficult. In this paper, an improved dual-phase-shifted 2-HB-LLC staggered resonant topology is proposed, which uses pair back-to-back active switches on secondary side to meet ultra-wide packs. switching frequency proposed identical frequency. Moreover, special structure with dual phase shift control strategy achieves high boost ratio,...