Data Analysis System Research Articles

Evaluation of changes caused by HR-HPV infection in squamous cell carcinoma of the head and neck using Raman microspectroscopy in combination with multivariate statistical analysis

Abstract Introduction: Squamous cell carcinoma of the head and neck region (HNSCC), with a positive status for high oncogenic potential human papillomavirus (HR-HPV), represents a clinically distinct disease entity compared to HPV-independent cases. Patients exhibit variations in prognosis and proposed therapy regimens. A prompt and reliable diagnosis of the presence of HPV infection could optimize the treatment for these patients. Currently employed treatment methods are long-term, expensive, and lack specificity, especially when administered separately. Material and methods: The research objective of this study is to explore significant differences in the Raman spectra of biological samples taken from patients with HNSCC, facilitating the confirmation of HPV virus presence. Study groups were delineated based on histopathological diagnosis and molecular biology tests, confirming the biological activity of the virus and the presence of the HR-HPV form with a diagnosis of a specific subtype. Results: To identify high oncogenic potential human papillomavirus (HR-HPV) infection as a crucial factor in squamous cell carcinoma of the head and neck region, an effective automatic data analysis system was established, relying on Raman microspectroscopy and multivariate analysis. Our results showed clear ranges of the Raman spectrum that differentiated between HPV-associated and non-HPV-associated cancers. Conclusions: In conclusion, our experience shows a great diagnostic potential of Raman confocal microscopy with multidimensional statistical analysis. In the future, the use of this method may allow for the creation of an effective and automated HR-HPV detection system in neoplastic tissue.

DEVELOPMENT OF THE INFORMATION PORTAL “KAZGEOWARNING” FOR FORECASTING FIRES ON THE TERRITORY OF KAZAKHSTAN

This article presents the development of an information system tailored for predicting and managing forest fires in Kazakhstan. The system's primary objective is to establish an information portal dedicated to fire forecasting, enhancing the accuracy of predictions and timely warnings. To achieve this, the system focuses on tasks such as predicting the likelihood of forest and steppe fires, monitoring environmental changes in real-time, and visualising the current situation on maps. It also generates comprehensive reports detailing fire risk, available resources, and necessary measures. Users will receive alerts about potential fire threats, facilitating rapid response. The system's data analysis incorporates geographical and climatic factors for precise forecasting. The ultimate goal is to significantly reduce the environmental and economic impact of forest fires in the country, emphasising the creation of a versatile platform for swift emergency response, integrating modern technologies into practical use. The practical significance encompasses improved fire warnings, enhanced safety for both the population and nature, and a reduction in economic losses. The accompanying website serves as a dynamic platform to collect, analyse, and visualise real-time data on disasters and crises, providing users with geographic information, maps, and reports crucial for making informed decisions during emergency situations.

Pre-operative lung ablation prediction using deep learning.

Microwave lung ablation (MWA) is a minimally invasive and inexpensive alternative cancer treatment for patients who are not candidates for surgery/radiotherapy. However, a major challenge for MWA is its relatively high tumor recurrence rates, due to incomplete treatment as a result of inaccurate planning. We introduce a patient-specific, deep-learning model to accurately predict post-treatment ablation zones toaidplanning and enable effective treatments. Our IRB-approved retrospective study consisted of ablations with a single applicator/burn/vendor between 01/2015 and 01/2019. The input data included pre-procedure computerized tomography (CT), ablation power/time, and applicator position. The ground truth ablation zone was segmented from follow-up CT post-treatment. Novel deformable image registration optimized for ablation scans and an applicator-centric co-ordinate system for data analysis were applied. Our prediction model was based on the U-net architecture. The registrations were evaluated using target registration error (TRE) and predictions using Bland-Altman plots, Dice co-efficient, precision, and recall, compared against the applicator vendor's estimates. The data included 113 unique ablations from 72 patients (median age 57, interquartile range (IQR) (49-67); 41 women). We obtained a TRE ≤ 2 mm on 52 ablations. Our prediction had no bias from ground truth ablation volumes (p = 0.169) unlike the vendor's estimate (p < 0.001) and had smaller limits of agreement (p < 0.001). An 11% improvement was achieved in the Dice score. The ability to account for patient-specific in-vivo anatomical effects due to vessels, chest wall, heart, lung boundaries, and fissures was shown. We demonstrated a patient-specific deep-learning model to predict the ablation treatment effect prior to the procedure, with the potential for improved planning, achievingcomplete treatments, and reduce tumor recurrence. Our method addresses the current lack of reliable tools to estimate ablation extents, required for ensuring successful ablation treatments. The potential clinical implications include improved treatment planning, ensuring complete treatments, and reducing tumor recurrence.

Virtual instrumentation-based data collection and analysis for CNC machining process

Using Computer Numerical Control (CNC) machine tools in manufacturing is crucial for exact processing materials. The quality of the machined product heavily relies on the machine's condition. Therefore, monitoring and assessing the machine's status is essential to ensure product quality and enhance its lifespan. However, due to cost, space, and technical limitations, only a few physical variables, such as acceleration, force, displacement, acoustic radiation, temperature, and flow conditions, can be measured. Efficient transfer and extraction of these data are critical for monitoring machine status using statistical methods or valuable signatures. Traditional measurement instruments are complex and numerous, whereas measurement and analysis systems based on virtual instrumentation technology collect necessary data from sensors and data acquisition cards, meeting the needs of test analysis. Virtual instrumentation utilizes computer hardware resources, modularization hardware, and software systems for data analysis, communication, and operation interface, providing greater versatility, flexibility, compatibility, and repeatability. Previous research has shown successful attempts at developing LabVIEW-based systems for monitoring CNC milling machines and analyzing cutting parameters and forces. This project aims to collect and process data from the CNC machining process. The main objectives include writing a LabVIEW software program, recording data from CNC machine programs using the provided hardware (myRIO) and LabVIEW program, processing the data using MATLAB software, and analyzing and discussing the obtained results. The report consists of four parts, starting with an introduction to the project's background and literature review, followed by a description of the project steps, methods, experiments, and data processing. The processed data graphs will be presented and discussed, and recommendations for future research will be provided before concluding with a study summary.

Design and Validation of a Stratified Shear Model Box for Seismic Response of a Sand-Blowing Reclamation Site

The global increase in building collapses and damage on soft-soil sites due to distant significant earthquakes poses similar challenges for sand-blowing reclamation (SBR) sites on soft-soil layers. This study was initiated to capture the vibration characteristics of the SBR sites and to provide fresh insights into their seismic responses. Initially, considering the heterogeneity and layered structure of soil at SBR sites, we developed a novel stratified shearing model box. This model box enables the simulation of the complex characteristics of soil layers at SBR sites under laboratory conditions, representing a significant innovation in this field. Subsequently, an innovative jack loading system was developed to apply active vertical pressure on the soil layer model, accelerating soil consolidation. Furthermore, a new data collection and analysis system was devised to monitor and record acceleration, pore water pressure, and displacement in real time during the experiments. To verify the model box’s accuracy and innovation, and to examine the seismic response of SBR sites under varying consolidation pressures, four vibration tests were conducted across different pressure gradients to analyze the model’s predominant period evolution due to consolidation pressures. The experimental results demonstrate that the model box accurately simulates the propagation of one-dimensional shear waves in soil layers under various consolidation pressures, with notable repeatability and reliability. Our experiments demonstrated that increasing consolidation pressure results in higher shear wave speeds in both sand and soft-soil layers, and shifts the site’s predominant period towards shorter durations. Concurrently, we established the relationship between the site’s predominant period and the input waves. This study opens new paths for further research into the dynamic response properties of SBR sites under diverse conditions through shaking-table tests.

Downscaling Future Precipitation over Mi Oya River Basin using Artificial Neural Networks

Studying future precipitation behaviour in river basins is essential for proper water resources and land-use planning within them, as this will help to reduce the risk and mitigate disasters that can occur in the future. General Circulation Models (GCMs) are used to study future precipitation fluctuations, which simulate large-scale climate variations under the effect of greenhouse gas changes. The GCM runs at a coarse spatial resolution which cannot be directly used for climate impact studies. Therefore, downscaling is required to extract the sub-grid and local scale information. This study examines the use of the Long Short-Term Memory (LSTM) neural network for climate downscaling to the Mi-Oya river basin in Sri Lanka using CNRM-CM5 and HadCM3 GCMs and observed annual data for 35 years. The precipitation data were extracted to cover Sri Lanka. Current downscaling models mostly use Convolutional Neural Networks (CNNs) to downscale GCMs. Out of 42 GCMs, two appropriate GCMs were chosen using the data analysis tool Data Integration and Analysis System (DIAS). The best predictor variables were chosen using the LASSO regression method. In this research, Machine Learning models were implemented using the Google TensorFlow platform. The Nash–Sutcliffe coefficient, Pearson correlation coefficient, and root-mean-square error performance indices were used to evaluate the performances of different downscaling models. Statistical downscaling was performed on the data at RCP 2.6, 4.5, and 8.5 using a LSTM. Subsequently, the changes that would take place by the year 2100 were analysed. The results show that precipitation will be reduced in the 2nd and 3rd decades of the 21st century, and precipitation will increase toward the 22nd century.

Cervical elastography using E-Cervix™ for prediction of preterm birth in twin pregnancies with threatened preterm labor

ObjectiveTo evaluate the prediction performance of E-Cervix™ for preterm birth in twin pregnancies with threatened preterm labor. MethodsThis was a single-center retrospective cohort study of twin pregnancies presenting to obstetrics triage for threatened preterm labor (PTL) between 23 0/7 – 33 6/7 weeks who received screening for PTL with transvaginal ultrasound cervical length (TVU CL) and cervical elastography with E-Cervix™ at the time of triage. Cervical elastography parameters were examined and compared between women who delivered preterm and those who did not. The quantification of cervical strain was calculated by a data analysis system that directly analyzes raw data from the region of interest (ROI) and described as hardness ratio (HR), mean strain level within 1 cm from internal (IOS) and external (EOS) os. Results63 twin gestations without prior preterm birth and with threatened PTL between 23 0/7 – 33 6/7 weeks of gestation were included in the study. 27 (42.9 %) had cervical length < 25 mm, and were admitted for true PLT. Out of the 36 women with cervical length ≥ 25 mm, 6 (16.7 %) were admitted. Women with threatened PTL had significantly higher HR compared to those with true PTL (p < 0.01), and significantly lower IOS and EOS. Women who delivered preterm had significantly higher HR compared to those who did not delivery preterm and significantly lower IOS and EOS, in overall cohort, and in the subset of women with true PTL. Incidences of HR < 50 % and < 35 % were statistically significantly higher in women who delivered preterm compared to those who did not (p < 0.01). ConclusionCervical elastography with E-Cervix™ may be useful for assessment of twin gestations presenting to obstetrics triage for threatened PTL.

Deep Learning Approaches for Medical Image Analysis and Diagnosis.

In addition to enhancing diagnostic accuracy, deep learning techniques offer the potential to streamline workflows, reduce interpretation time, and ultimately improve patient outcomes. The scalability and adaptability of deep learning algorithms enable their deployment across diverse clinical settings, ranging from radiology departments to point-of-care facilities. Furthermore, ongoing research efforts focus on addressing the challenges of data heterogeneity, model interpretability, and regulatory compliance, paving the way forseamless integration of deep learning solutions into routine clinical practice. As the field continues to evolve, collaborations between clinicians, data scientists, and industry stakeholders will be paramount in harnessing the full potential of deep learning for advancing medical image analysis and diagnosis. Furthermore, the integration of deep learning algorithms with other technologies, including natural language processing and computer vision, may foster multimodal medical data analysis and clinical decision support systems to improve patient care. The future of deep learning in medical image analysis and diagnosis is promising. With each success and advancement, this technology is getting closer to being leveraged for medical purposes. Beyond medical image analysis, patient care pathways like multimodal imaging, imaging genomics, and intelligent operating rooms or intensive care units can benefitfrom deep learning models.

Основные аспекты цифрового развития общества

The digital development of society is an urgent and important topic in the modern world, as technologies and digital tools play an increasingly important role in our daily lives. This research topic is an urgent phenomenon that requires attention and study for the effective implementation of digital technologies in various spheres of life. The purpose of the work is to study the main aspects of the digital development of society, their use and impact on people’s lives. To achieve the goal, the following were studied: the development of digital technologies and devices, the creation of digital infrastructure, the use of digital technologies to increase productivity and efficiency in various fields, the use of digital technologies to ensure a decent human life, as well as opportunities and risks. Thus, the relevance of this work lies in describing the impact of digitalization on all processes taking place in society as a whole. The paper reflects the directions of the use of the Internet and the use of digital technologies in organizations, identifies the share of the use of digital technologies to increase productivity and efficiency in agriculture and higher education, notes the development of electronic services for public services in the field of public administration and social services, improving the efficiency of the health system, the introduction of electronic voting and digital identification, the development of data monitoring and analysis systems and more. It is noted that along with the possibilities of digitalization, risks arise. Thus, digitalization has affected the economy, education, healthcare, government, cultural and entertainment development, and security.

Enhancing Study Space Utilization at UCL: Leveraging IoT Data and Machine Learning

University College London (UCL) has more than 51,000 students along with a challenge to find available study space. UCL has improved its Internet of Things (IoT) infrastructure to better analyze space occupancy in order to solve this. Current systems offer real-time information, but they are only available in crowded places. Therefore, this study presents a data analysis system that recommends the best study areas by using Machine Learning (ML) and historical data. The main library, student center, and science library at UCL are the top three study places. The goal of this study is to transform raw occupancy data from these venues into useful insights. The development of an independent system written in Python was facilitated by research into dataset optimization, ML regressions, popularity patterns, and feature analysis. This technique uses historical trends to forecast occupancy rates for the future.

Heterogeneous Internet of Things Big Data Analysis System Based on Mobile Edge Computing

The big data heterogeneous Internet of Things (IoT) requires mobile edge computing (MEC) to process some data, and the data analysis system of MEC often has the problem of excessive terminal energy consumption (ECS) or long delay. So this study designed an energy-saving optimization algorithm for the task offloading processing module in the big data heterogeneous IoT analysis system, and designed and conducted simulation experiments to verify the application performance of the algorithm. The experimental results show that the #04 scheme of the designed algorithm has the lowest terminal ECS under the same conditions. Choosing the #04 scheme to build the algorithm, comparative analysis shows that when the edge server (ES) computing rate is 10 cycles/s, the weighted sum values of terminal ECS for EOPU, MPCO, exhaustive search, and local computing methods are 23.6 J, 23.9 J, 28.5 J and 84.5 J, respectively. Moreover, the algorithm possesses a significantly higher percentage of remaining time under different conditions of total SMD devices and total subchannels compared to other methods. This indicates that the designed algorithm can markedly enhance the processing performance of the task offloading model of the big data heterogeneous IoT data analysis system, and can also effectively reduce terminal ECS and system latency. The research results can provide reference for improving the processing ability of heterogeneous IoT big data analysis systems. The contribution of this study to the academic field lies in providing a model that can effectively reduce the operational ECS and time consumption of heterogeneous IoT big data analysis systems containing mobile animal networking devices. Moreover, from an industrial perspective, the results of this study contribute to improving the efficiency of information exchange and processing in the field of IoT computing, thereby promoting the promotion of IoT technology.

Data mining algorithm in the identification of accounting fraud by smart city information technology

The identification of accounting fraud is an important measure to safeguard the interests of stakeholders and ensure the long-term development of the company. The current traditional methods for identifying accounting fraud rely on manual review and judgment, lacking objectivity and accuracy. In order to improve the accuracy of accounting fraud identification, improve identification efficiency and objectivity, this article combines smart city information technology to conduct in-depth research on data mining algorithms for accounting fraud identification. This article first provides a brief overview of smart cities and information technology, then introduces the basic theory of accounting fraud identification, and finally implements accounting fraud identification through k-means clustering mining algorithm. The data is divided into k clusters, and abnormal clusters are identified by checking the characteristics and attributes of each cluster. Compared with traditional rule-based and pattern based methods, this approach can more flexibly adapt to different types and forms of fraud, and can discover unknown patterns of fraud. In the experiment, this article used electronic data collection, analysis, and retrieval systems on the websites of the Shanghai Stock Exchange and Shenzhen Stock Exchange to collect 641 annual reports and financial characteristics from 62 listed companies that engaged in financial statement fraud and 84 companies that were not reported to have financial statement fraud from 2012 to 2021 as test samples. The results were tested and analyzed from several aspects, including the number of misjudgments, misjudgment rate, and ROC curve. The final test results show that compared to traditional accounting fraud identification methods, the comprehensive misjudgment rate of data mining algorithms based on smart cities has decreased by 3 %. The conclusion indicates that data mining algorithms used in smart city information technology to identify accounting fraud can help improve the accuracy of accounting fraud, improve audit objectivity and effectiveness.

Development of a modular structural geographic information system for spatial data analysis

This work contains the research on the development and implementation of a modular structural geographic information system (GIS) for spatial data analysis, which is based on artificial intelligence technologies, mainly machine learning. The article discusses the basic principles and concepts of modularity, and also proposes a methodology for developing a modular GIS structure using modern technologies and tools. The study has practical significance for specialists in the field of geoinformatics and spatial analysis, and can also be useful for GIS developers and scientific researchers interested in creating modular GIS structures for analyzing spatial data.

GEODESY IN SOLVING PROBLEMS OF SUSTAINABLE CITY DEVELOPMENT: ANALYSIS OF URBAN TERRITORY PLANNING

The article focuses on the creation and implementation of a geodetic planning system aimed at ensuring the sustainable development of urban areas. The main emphasis is on the use of open and accessible software, integration of full-fledged GIS and DBMS components. The process of determining the most suitable GIS platform for GSP included a detailed analysis of various GIS systems taking into account their functionality, which allowed to choose QGIS as the optimal solution due to its high functionality in the field of geodata management, support for various data formats, user interface and dynamic development. The structure of the GSP is based on two main platforms: the QGIS GIS system for cartographic data presentation and analysis, and the Microsoft Access DBMS for performing all data operations. Interaction between these platforms is provided through the eVis module built into QGIS, which allows for efficient planning of urban area surveys and land use management. The main tasks solved with the help of GSP are the planning of primary surveys to identify target plantations, detailed taxation of plantations with the creation of samples according to specified parameters, obtaining point maps of the distribution of various types of vegetation, planning sample surveys with specific geospatial conditions, as well as geo-information and tax analysis in in solving the problems of sustainable development of cities. The implementation of GSP provides significant advantages for planning the sustainable development of urban areas, allowing more efficient use of land resources, optimizing urban management processes and intensifying scientific research in this field. The development and integration of modern digital technologies in the field of urban planning play a key role in achieving the goals of sustainable development.

Detecting Halyomorpha halys using a low-power edge-based monitoring system

Smart monitoring systems in orchards can automate agriculture monitoring processes and provide useful information about the presence of insects, such as the Brown Marmorated Stink Bug (BMSB), that threaten the production quantity and quality of fruit such as pears. Unlike other approaches in the literature, we propose a low-cost image monitoring system which exhibits a very low power consumption without compromising much of the accuracy that existing expensive systems incorporating significant computing and processing capability can achieve in such applications. The proposed system relies on a microcontroller unit and a camera which can take pictures of a double-sided sticky insect trap which, with the help of novel machine learning algorithms, can report on the presence of BMSB via a long-range communication link. The Internet of Things data capture and analysis system has recently been deployed in a real orchard in Italy which is subject to BMSB infestation and the first images have been analysed. This paper presents how the system works, the image processing, detection and classification algorithms, as well as a demonstration of the memory and energy consumption associated with the processing algorithms. The system achieves an accuracy of over 90% with multiple times less memory and energy consumption compared to other similar approaches in the literature.

Developing excellent plantar pressure sensors for monitoring human motions by using highly compressible and resilient PMMA conductive iongels

Based on real-time detection of plantar pressure, gait recognition could provide important health information for rehabilitation administration, fatigue prevention, and sports training assessment. So far, such researches are extremely limited due to lacking of reliable, stable and comfortable plantar pressure sensors. Herein, a strategy for preparing high compression strength and resilience conductive iongels has been proposed by implanting physically entangled polymer chains with covalently cross-linked networks. The resulting iongels have excellent mechanical properties including nice compliance (young’s modulus < 300 kPa), high compression strength (>10 MPa at a strain of 90 %), and good resilience (self-recovery within seconds). And capacitive pressure sensor composed by them possesses excellent sensitivity, good linear response even under very small stress (∼kPa), and long-term durability (cycles > 100,000) under high-stress conditions (133 kPa). Then, capacitive pressure sensor arrays have been prepared for high-precision detection of plantar pressure spatial distribution, which also exhibit excellent sensing performances and long-term stability. Further, an extremely sensitive and fast response plantar pressure monitoring system has been designed for monitoring plantar pressure of foot at different postures including upright, forward and backward. The system achieves real-time tracking and monitoring of changes of plantar pressure during different static and dynamic posture processes. And the characteristics of plantar pressure information can be digitally and photography displayed. Finally, we propose an intelligent framework for real-time detection of plantar pressure by combining electronic insoles with data analysis system, which presents excellent applications in sport trainings and safety precautions.

Development of an intelligent system for doctor-patient interaction in a day hospital

The article discusses the issue of finding optimal digital solutions to streamline organizational processes for medical organizations when providing medical care in a day hospital setting. Based on the analysis of in-depth interviews with medical specialists, the specific features of the system of providing medical care in a day hospital setting are described, and areas for optimizing the procedure for shortterm hospitalization at the present stage are identified. Taking into account the needs of practical healthcare in digital technologies, an algorithm for the functioning of a novel intellectual product (decision support system based on the intellectual processing method) is presented, including an onpremise system for monitoring, diagnosis, and analysis of medical data (condition self-assessment, satisfaction with aspects of treatment, etc.).

Landsat 9 Transfer to Orbit of Pre-Launch Absolute Calibration of Operational Land Imager (OLI)

Landsat 9 Operational Land Imager (L9-OLI) was launched on 27 September 2021, after completing a successful radiometric pre-launch calibration and characterization phase. The radiometric math model that governs the ground system—the data processing and analysis system (DPAS)—uses various calibration parameters that had been derived based on the pre-launch tests and analysis. During the on-orbit commissioning phase, the OLI system acquired specific sets of data collects, which enabled the revalidation of the pre-launch absolute calibration scale and other associated instrument performance characteristics. The analysis results shown in this paper focus on the activities and results related to the transfer-to-orbit analysis for the SI-traceable pre-launch radiometric scale. Key topics discussed in this paper include: radiance and reflectance calibration parameters for OLI; solar diffuser collects; stimulation-lamp collects; dark response; signal-to-noise ratios; and noise characteristics; radiometric response stability and the on-orbit update to the radiance to reflectance conversion factors. It will be shown that the OLI response during the early on-orbit operation matched pre-launch results and therefore this re-validates the absolute radiometric scaling at the predicted pre-launch level within the expected level of uncertainties. The launch did not cause any significant changes to the OLI system from the perspective of the absolute radiometric calibration performance. Once the transfer to orbit of the absolute calibration was confirmed, it created a solid basis for further on-orbit refinements of the radiance calibration parameters. As such, follow-on calibration refinements are discussed in other articles within this special issue, and they address issues such as uniformity as well as cross-calibration activities.

Hospital Medical Behaviour Supervision and Operational Efficiency Evaluation Method based based on Big Data Platform

The application of cloud computing and big data core technologies and concepts to medical informatization can improve its flexibility and efficiency, and achieve the overall deployment and intensive management of the system. In the era of big data, the use of information management platform can optimize and standardize clinical diagnosis and treatment process, improve the quality and efficiency of diagnosis and treatment services, and improve the quality and level of scientific research, which meets the requirements of medical reform for fine management of hospitals and precise medical treatment in the era of evidence-based medicine. This paper introduces the development and application of big data analysis in the field of information technology. Through the research on the supervision of hospital medical service behavior, the top-level design is used to standardize the content of medical behavior supervision, and the effectiveness of supervision is discussed to achieve the purpose of reducing clinical paperwork. Based on the needs, the medical service behavior supervision system is proposed and constructed. Strengthen hospital medical behavior supervision through hospital big data analysis and knowledge base system support. This system can provide management with an effective monitoring and management tool, enabling them to promptly identify and solve problems that arise in medical services. By analyzing a large amount of medical data, hospitals can help predict the trend of disease occurrence in advance, thereby taking preventive and control measures in advance, and reducing the occurrence and spread of diseases. Through deep learning and analysis of patient data, personalized treatment plans can be provided for each patient to improve treatment effectiveness.

Remote Intelligent Medical Monitoring Data Transmission Network Optimization based on Deep Learning

A hospital operating status evaluation data analysis system was established based on the autoencoder's network. The Gibbs sampling method is used to obtain the approximate distribution of RBM. In addition, the Autoencoder neural network can also select feature dimensions that can better characterize the characteristics of financial operation data from a large amount of financial operation data. Deep learning methods are used to study the redundant information elimination method and the generation mechanism of multi-source heterogeneity in multi-source heterogeneous networks. The principle of intrinsic compression is used to reduce the dimensionality of the redundancy in the network and obtain the compression redundancy objective function. This article sets thresholds for information classification on the Internet. The approach was tested using financial data from a medical institution. Use smart encoders to extract 17 financial indicators from financial data that can be used for modeling. The evaluation results are used as the output vector of the model. Comparative experiments show that the AUC value and accuracy of the method proposed in this article can be improved by 0.84 and 83.33% compared with the AUC value of shallow logistic regression and BP neural network. This algorithm has apparent improvements.