Research on Optimization Algorithm for Knowledge Graph Construction in Education Quality Evaluation

With the rapid development of smart grid technology, a large amount of multi-source heterogeneous data has been generated in the power system, and its effective utilization is crucial for the optimization operation, demand prediction, and anomaly detection of the power system. However, the fusion processing of multi-source heterogeneous data faces many challenges, such as inconsistent data format, granularity, and quality, and direct fusion can easily lead to information redundancy and contradictions. A multi-source heterogeneous data fusion technology based on big data mining has been proposed to address the above issues. This method combines the advantages of convolutional neural networks and gated recurrent units to automatically extract features from image and sequence data and handle long-term dependency issues in time series data. Meanwhile, the K-means clustering algorithm is used to preprocess the data and train a specialized ConvGRU model. The results showed that in short-term load forecasting and abnormal electricity consumption behaviour detection tasks, the accuracy of this method reached 96.3% and 98.7%, respectively, with AUC values of 0.994 and 0.996. Compared to models that use only CNN or GRU, the performance is significantly improved. This method effectively solves the problem of integrating and processing multi-source heterogeneous power data, improves the accuracy and efficiency of power system data analysis, and provides strong support for the optimized operation of smart grids.

The study of the prediction of stock market volatility is of great significance to rationally control financial market risks and increase excessive investment returns and has received extensive attention from academic and commercial circles. However, as a dynamic and complex system, the stock market is affected by multiple factors and has a comprehensive capability to include complex financial data. Given that the explanatory variables of influencing factors are diverse, heterogeneous and complex, the existing intelligent algorithms have great limitations for the analysis and processing of multi-source heterogeneous data in the stock market. Therefore, this study adopts the edge weight and information transmission mechanism suitable for subgraph data to complete node screening, the gate recurrent unit (GRU) and long short-term memory (LSTM) to aggregate subgraph nodes. The compiled data contain the metapaths of three types of index data, and the introduction of the association relationship attention dimension effectively mines the implicit meanings of multi-source heterogeneous data. The metapath attention mechanism is combined with a graph neural network to complete the classification of multi-source heterogeneous graph data, by which the prediction of stock market volatility is realized. The results show that the above method is feasible for the fusion of heterogeneous stock market data and the mining of implicit semantic information of association relations. The accuracy of the proposed method for the prediction of stock market volatility in this study is 16.64% higher than that of the dimensional reduction index and 14.48% higher than that of other methods for the fusion and prediction of heterogeneous data using the same model.

As the complexity of intelligent public transportation systems continues to increase and the degree of interconnection of equipment at all levels gradually deepens, building a digital twin system will generate a large amount of multi-source heterogeneous data. Effective processing and in-depth mining of data can provide real-time and intelligent decision-making optimization for the construction of digital twin systems, thereby efficiently handling complex emergencies. This paper conducts a systematic review on the processing methods and technologies of multi-source heterogeneous twin data in traffic dynamic scenarios. First, the content and classification of multi-source heterogeneous twin data in the operation process of intelligent bus systems are clarified; secondly, the multi-source The data processing methods and technologies applied in each stage of data collection, data storage, data fusion and data analysis in heterogeneous twin data processing are analyzed, and the advantages, disadvantages and applications of various methods and technologies are analyzed; finally, corresponding to the multiple This paper summarizes the methods and technologies for processing multi-source heterogeneous twin data, and points out the challenges and development trends faced by the current multi-source heterogeneous twin data processing methods and technologies.

In this paper, a multi-scale forecasting model of wind power in extreme weather based on multi-source heterogeneous data fusion is studied. The intermittence and fluctuation of wind power, especially the frequent occurrence of extreme weather events, significantly increase the uncertainty of wind power forecasting and bring challenges to the stable operation of power system. Traditional forecasting methods rely on a single data source, so it is difficult to fully describe the coupling relationship between complex meteorological conditions and wind farm dynamics, and there are limitations in extreme weather scenarios. In this study, a multi-scale prediction model combining weather radar, satellite cloud images and SCADA data is proposed. Through time-space alignment and feature extraction, extreme weather enhancement module and hierarchical multi-scale prediction model, the problems of multi-modal information integration, sensitivity enhancement of extreme weather events and multi-scale time series prediction are solved. The experimental results show that the prediction accuracy of this model is significantly better than the existing models in extreme weather such as typhoon, and it shows a good balance in calculation efficiency, which provides effective support for ensuring power grid security, optimizing energy storage configuration and reducing wind abandonment rate.

Based on the multi-source and heterogeneous characteristics of massive state sensing data of bridges, this paper establishes a multi-source heterogeneous bridge state sensing data feature extraction and analysis method. A dynamic evaluation model of bridge technology based on multi-source heterogeneous data fusion is proposed. For the first time, the relationship between multi-source perceptual data and bridge infrastructure technology is established, the knowledge graph of bridge infrastructure technology is constructed, and the relationship between various state-aware data and bridge technology is explicitly expressed by the map. This collaborative association is used to dynamically evaluate and predict bridge technology.

In recent years, the demand for urban travel is increasing and the travel modes are diverse. Online car Hailing has become an important way to meet the travel needs of residents. The online car-hailing platform receives tens of thousands of travel requests every day. However, a large portion of the thousands of orders are unfinished, that is, canceled by passengers. This not only reduces the income of drivers but also affects the order dispatching efficiency of the online car-hailing platform. To predict the cancellation probability of online car-hailing orders(OCP), the relationship between multi-source heterogeneous data and OCP is first introduced, in which the presence of idle taxis is the main factor for passengers to cancel their orders during the waiting period. Secondly, a deep learning model based on the Seq2Seq structure is designed to predict OCP in real-time. The model consists of an attribute fusion module, encoder layer, and decoder layer. Finally, a full experiment is carried out using the Didi Chengdu online car-hailing order data set to verify the effectiveness of the algorithm.KeywordsTaxi order cancellationUrban travelFeature fusionDeep learning

The complexity and diversity of equipments presented in the development of modern industrial technology brings the wide application of deep learning in fault diagnosis. However, the multi-source heterogeneous data collected from different sensors applied in industrial production cannot be used by traditional neural networks directly, which registers a difficulty to fault diagnosis based on deep learning. In addition, with only one of the data used, traditional deep learning fault diagnosis methods ignores the correlation between heterogeneous data, which may result in the loss of useful information and the accuracy of fault diagnosis. To solve the above problems, a fault diagnosis framework based on deep learning for multi-source heterogeneous data fusion is proposed. In the feature learning stage, stack self-encoder (SAE) and convolution neural network (CNN) are used to extract deep features from one-dimensional vibration data and two-dimensional image data respectively, and then feature fusion fault diagnosis is carried out. It solves the problem that traditional neural network can't make full use of multi-source heterogeneous data, and improves the accuracy of fault diagnosis. What's more, the bearing data of Case Western Reserve University are used to verify the validity of the diagnosis method proposed in this paper.

Traditional QA systems often rely on shallow semantic retrieval and single-source knowledge, making them insufficient for handling cross-domain queries and multi-hop reasoning. To address these limitations, this study proposes a novel QA framework that integrates multi-source heterogeneous data fusion with knowledge graph-based reasoning. The framework first introduces a multi-channel alignment and fusion module that unifies heterogeneous knowledge sources through entity alignment, relation extraction, and confidence-weighted attribute fusion, constructing a coherent knowledge graph. On this basis, a graph augmentation strategy enhances structural connectivity and semantic completeness. A deep reasoning module, combining relation-aware graph neural networks with query-conditioned multi-hop attention, enables both accurate inference and interpretable reasoning paths. Finally, a joint decoder simultaneously outputs natural language answers and supporting reasoning chains, improving explainability. Extensive experiments conducted on WikiHop, HotpotQA, and ComplexWebQuestions demonstrate the superiority of our model. Compared to strong baselines such as GRAFT-Net, HGN, and KG-QA-Base, our framework achieves consistent improvements of 4–6% in Exact Match (EM) and 5–7% in F1 score. More importantly, Path-Precision improves by 10–12 percentage points, highlighting the ability to generate reliable reasoning evidence. This research not only provides a feasible technical pathway for building interpretable QA systems with deep understanding, but also carries important theoretical and practical implications for advancing the cognitive capabilities of general artificial intelligence.

In medical domain, huge amounts of data are generated at all times. These data are usually difficult to access, with poor data quality and many data islands. Besides, with a wide range of sources and complex structure, these data contain essential information and are difficult to manage. However, few existing data management frameworks based on Data Lake excel in solving the persistence and the analysis efficiency for medical multi-source heterogeneous data. In this paper, we propose an efficient Multi-source Heterogeneous Data Lake Platform (MHDP) to realize the efficient medical data management. Firstly, we propose an efficient and unified method based on Data Lake to store data of different types and different sources persistently. Secondly, based on the unified data store, an efficient multi-source heterogeneous data fusion is implemented to effectively manage data. Finally, an efficient data query strategy is carried out to assist doctors in medical decision-making. In-depth analysis on applications shows that MHDP delivers better performance for data management in medical domain.

A multi-source heterogeneous spatial big data fusion method based on multiple similarity and voting decision

With the growing importance of the fuzzy spatiotemporal data in information application, there is an increasing need for researching on the integration method of multi-source heterogeneous fuzzy spatiotemporal data. In this paper, we first propose a fuzzy spatiotemporal RDF graph model based on RDF (Resource Description Framework) that proposed by the World Wide Web Consortium (W3C) to represent data in triples (subject, predicate, object). Secondly, we analyze and classify the related heterogeneous problems of multi-source heterogeneous fuzzy spatiotemporal data, and use the fuzzy spatiotemporal RDF graph model to define the corresponding rules to solve these heterogeneous problems. In addition, based on the characteristics of RDF triples, we analyze the heterogeneous problem of multi-source heterogeneous fuzzy spatiotemporal data integration in RDF triples, and provide the integration methods FRDFG in this paper. Finally, we report our experiments results to validate our approach and show its significant superiority.

Aiming at the low cleaning rate of the traditional multi-source heterogeneous power grid big data cleaning model, a multi-source heterogeneous power grid big data cleaning model based on machine learning classification algorithm is designed. By capturing high-quality multi-source heterogeneous power grid big data, weight labeling of data source importance measurement, data attributes and tuples, and constructing Tan network based on the idea of machine learning classification algorithm, the data probability value is finally used to complete the classification and cleaning of inaccurate data. Experiments show that the model based on machine learning classification algorithm can effectively improve the imprecise data cleaning rate compared with the traditional model to solve multi-source heterogeneous imprecise data cleaning.

In order to fully utilize the features of multi-source heterogeneous data and effectively improve the accuracy and efficiency of fault diagnosis of rotating machinery, a multi-source heterogeneous data diagnosis method based on parameter collaborative optimization multi-scale convolutional autoencoder (MSCAE) is proposed. Firstly, multi-scale information learning is integrated into the convolutional autoencoder (CAE) to consider the temporal and spatial feature information of the diagnostic object simultaneously. To improve the training and diagnostic efficiency of MSCAE, a quantum particle swarm optimization (QPSO) module is used to perform hyperparameter optimization on it using chaos initialization and dynamic weight strategy (DWS). Besides, the sparse attention mechanism is introduced into the MSCAE model to improve the recognition rate of key fault features hidden in the original heterogeneous signals. Finally, the confusion matrix and visualization techniques are used to achieve fault classification. The experimental results demonstrate that after 100 experiments, the proposed method has an average diagnostic accuracy of 98.5 % and strong robustness to noise, providing a new method for rotating machinery fault diagnosis based on multi-source heterogeneous data.

With the continuous development of aviation technology, the complexity of aircraft structures continues to increase, and higher requirements are placed on their strength test verification. Traditional test verification methods can no longer meet the requirements of new aircraft structures for high precision, real-time performance, and data fusion. This paper studies the application of cyber-physical fusion technology in aircraft structural strength testing, focusing on key technologies such as PINN-based virtual model calibration and improvement, and multi-source heterogeneous data fusion. Through these methods, stress field reconstruction, sensor data fusion noise reduction, and dynamic response display under limited measurement point conditions are achieved, thereby significantly improving the accuracy and efficiency of structural strength verification. The research results provide new technical support for structural reliability assessment, early fault diagnosis, and health management, and contribute to proactive maintenance and life cycle management.

In order to improve the prediction accuracy, this paper proposes a traffic accident risk prediction method of tunnel based on multi-source heterogeneous data fusion. Firstly, the feature extraction and coding model based on Gabor cloud image is constructed, and the unstructured cloud image is enhanced by amplitude feature level and then encoded by fusion. Secondly, the long sequence formed after the sample concatenation of the multi-source heterogeneous data of tunnel is used as the input of the prediction model GRU (Gate Recurrent Unit), and AdaBoost (Adaptive Boosting) is introduced to learn the prediction results to further improve the robustness of the prediction model. The experimental results show that the proposed GRU-AdaBoost model achieves the prediction accuracy of 84.59% when the data volume is 3 years and the time interval is 5 weeks. When the data volume is 1 year and the time interval is 3 weeks, the prediction accuracy can reach 80.85%, which is 9.61% higher than traditional prediction models.