Many state of the art methods analyze sentiments in news to predict stock price. When predicting price movement, correlation between stocks is a factor that can't be ignored because correlated could cause co-movement. Traditional measuring are mostly based on similarity corresponding data, while ignoring business relationships companies, such as shareholding, cooperation and supply-customer relationships. To solve this problem, paper proposes new method calculate by using enterprise...

In the realm of federated learning (FL), exchange model data may inadvertently expose sensitive information participants, leading to significant privacy concerns. Existing FL privacy-preserving techniques, such as differential (DP) and secure multi-party computing (SMC), though offering viable solutions, face practical challenges including reduced performance complex implementations. To overcome these hurdles, we propose a novel pragmatic approach preservation in by employing localized...

There are several unsolved problems in federated learning, such as the security concerns and communication costs associated with it. Differential privacy (DP) offers effective protection by introducing noise to parameters based on rigorous definitions. However, excessive addition can potentially compromise accuracy of model. Another challenge learning is issue high costs. Training large-scale models be slow expensive terms resources. To address this, various model pruning algorithms have...

The aims of our research are to evaluate the prediction performance proposed neuro-fuzzy model with System Evaluation and Estimation Resource Software Model (SEER-SEM) in software estimation practices apply architecture that combines technique different algorithmic models. In this paper, an approach combining SEER-SEM effort algorithm is described. This possesses positive characteristics such as learning ability, decreased sensitivity, effective generalization, knowledge integration for...

Software effort estimation is a critical part of software engineering. Although many techniques and algorithmic models have been developed implemented by practitioners, accurate prediction still challenging endeavor. In order to address this issue, novel soft computing framework was previously developed. Our study utilizes develop an approach combining the neuro-fuzzy technique with System Evaluation Estimation Resource - Model (SEER-SEM). Moreover, our assesses performance proposed model...

To trace the copyright of deep neural networks, an owner can embed its identity information into model as a watermark. The capacity watermark quantify maximal volume that be verified from watermarked model. Current studies on focus ownership verification accuracy under ordinary removal attacks and fail to capture relationship between robustness fidelity. This paper network watermarks theoretical perspective. We propose new definition analogous channel capacity, analyze properties, design...

Abstract：Some problems are existed in confirming the weighted value software quality model, for example, method is relatively rough and how abnormal rejected while judging value.Aiming at these problems, it pointed out that scientific hierarchy structure of characteristics sub should be built based on analytic process.Through construction judgment matrix, single hierarchical arrangement, total arrangement consistency check, comprehensive bottom will confirmed.Then by extreme rejecting...

Accurate software development effort estimation is critical to the success of projects. Although many techniques and algorithmic models have been developed implemented by practitioners, accurate prediction still a challenging endeavor in field engineering, especially handling uncertain imprecise inputs collinear characteristics. In this paper, hybrid in-telligent model combining neural network integrated with fuzzy (neuro-fuzzy model) has used improve accuracy estimating cost. The...