286results about How to "Precise screening" patented technology

The invention relates to the technical field of infrared image recognition, and discloses a hot target extraction method and system based on infrared image recognition, and the system comprises a multi-sensor collection module, a data dimension reduction module, a data verification module, a segmentation optimization module, and a dynamic feedback module. Multi-modal data dimension reduction processing is achieved by collecting infrared images, visible light images and environment dynamic parameters and combining Laplacian feature mapping, and a target area is extracted through geometric verification and topological analysis. An energy optimization model and a distribution alignment method are adopted to complete accurate segmentation of a target area, and meanwhile, segmentation model parameters are adjusted in real time based on environment dynamic parameters, so that the system stability is ensured. According to the method, high-precision and robust extraction of the thermal target is realized in a complex dynamic environment, the problems of fuzzy target boundary, noise interference and insufficient dynamic environment adaptability in the prior art are effectively solved, and the method is suitable for the fields of military reconnaissance, fireproof monitoring, electric power inspection and the like.

The invention relates to the technical field of carrier communication, in particular to a multichannel power line carrier communication system which comprises a data acquisition module, a first determination module, a second determination module, a judgment module, a selection module, a sending module and a correction module. According to the invention, through an accurate data acquisition and real-time determination mechanism, complex environment changes in a power network can be effectively dealt with, stability and high efficiency of communication are ensured, channels with optimal parameters such as signal strength, noise, delay and the like are optimized through multi-stage screening and determination, channels with relatively large interference and noise are effectively avoided, and the communication efficiency is improved. Meanwhile, the correction module can adjust the intensity threshold value in real time, it is guaranteed that the system still maintains excellent communication performance when the number of channels fluctuates and the network load changes, and the problems that due to the fact that the attention degree of the influence of real-time environment changes on signal quality is low, channel selection is inaccurate, and verification results are deviated are effectively solved.

The invention relates to the technical field of knowledge base construction, in particular to a knowledge service system construction method, which comprises a knowledge acquisition step for collecting knowledge data from knowledge sources of multiple fields; the preprocessing step is used for preprocessing the knowledge data; the classification step is used for classifying the obtained knowledge data; the similarity calculation step is used for calculating the similarity between different categories of knowledge bases; the first judgment step is used for judging whether the similarity is greater than a first threshold value or not; the association step is used for establishing association between the knowledge according to the similarity; the second judgment step is used for judging whether the similarity is greater than a second threshold value or not; the labeling step is used for constructing a new upper label; a label directory generation step for establishing a label directory; a query demand acquisition step for acquiring a knowledge retrieval demand of a user and generating a query request; the knowledge retrieval step is used for conducting knowledge recommendation according to the query request. The method has the effect of improving the searching efficiency.

The invention belongs to the technical field of low-voltage distribution cable fault identification, and particularly relates to a method for judging a series arc fault of a low-voltage distribution cable based on disturbance current characteristics, which comprises the following steps of: S1, acquiring current data of three-phase current of the cable; s2, analyzing the similarity of the current waveforms of the adjacent current periods by adopting a Pearson correlation coefficient PCC, and judging whether the cable has current disturbance or not; s3, when it is judged that current disturbance occurs, intercepting current data of a disturbance starting point and a plurality of current periods after the disturbance starting point, and recording the current data as disturbance current; calculating the time-frequency domain characteristic quantity of the intercepted disturbance current; and S4, inputting the time-frequency domain characteristic quantity of the disturbance current into a preset fault identification model, carrying out detection and identification on a cable series arc fault, and judging whether the fault causing current disturbance is the series arc fault or not. According to the method, the low-voltage distribution cable series arc fault can be efficiently, accurately and timely monitored and identified, and the safety and stability of a power system are improved.

The invention discloses an agent task scheduling planning method, and relates to the technical field of agent scheduling. The method comprises the steps of analyzing a task instruction to generate atomic tasks capable of being independently executed, constructing a subtask dependency graph, and defining association constraints between the tasks; determining the real-time resource occupancy state of the intelligent agent to obtain a resource state tensor, completing resource-task association anchoring and dependency priority ranking in combination with the sub-task dependency graph, and generating a task priority sequence with resource constraint; performing dynamic capability matching and predictive load balancing calculation on the sequence through a task-agent adaptation model, and determining a target execution agent of each atomic task; and based on the target execution agent and the subtask dependency graph, performing time sequence scheduling arrangement and conflict resolution, and generating a collaborative execution scheme. The method improves the reasonability and efficiency of agent task scheduling, reduces resource conflicts and execution timeout risks, and is suitable for agent cluster collaborative scheduling in a complex scene.

The invention discloses a vehicle visual relay tracking method based on spatial-temporal feature fusion, and belongs to the technical field of intelligent traffic. After a multi-target tracking trajectory of each camera on a road is obtained, the method comprises the following steps: establishing a camera space-time constraint based on a topological relation between the cameras and the road, performing trajectory compatibility judgment based on the camera space-time constraint, and filtering out incompatible trajectory pairs; modeling the trajectory of the target vehicle into a graph, and calculating the similarity of compatible trajectory pairs based on a graph convolutional neural network; based on the trajectory similarity, the multi-camera trajectory association problem is decomposed, hierarchical clustering is performed on the trajectory by using a heap data structure, and the global association trajectory of each target vehicle is obtained step by step. According to the method, the efficiency and accuracy of track association are improved, the calculation efficiency is optimized, and vehicle tracking in a complex scene and under multiple cameras can be dealt with.

The invention discloses a large model reasoning method and device based on sensitive information protection and related equipment. The method comprises the steps that firstly, an original input text is obtained, sensitive entities in the original input text are detected through a named entity recognition NER model, relation extraction and attribute extraction are conducted, and a sensitive information knowledge graph is constructed; then, screening sensitive information in the original input text according to the knowledge graph, and replacing the sensitive information to generate a desensitized input text; next, calling a first large model of the server side, and executing a specified task by using the desensitized input text to obtain a result text; and finally, calling a second large model of the user side, and restoring the sensitive information in the result text by using the original input text, the desensitized input text and the result text to obtain a complete text. Large models configured at different ends are utilized, sensitive information is protected, and meanwhile practicability of result texts is taken into account.

The invention belongs to the field of drug design, and particularly discloses a virtual screening method of a small-molecule inhibitor targeting YTHDF1 protein, which comprises the following steps: acquiring a YTHDF1 protein structure, and carrying out homologous modeling to generate various conformation proteins; small molecules combined with the YTHDF1 protein are obtained, bait molecules are generated, and ligand molecules are formed; a ligand molecule is in butt joint with a binding site on the YTHDF1, and the compound is characterized through PLEC fingerprints; the method comprises the following steps: constructing a specificity scoring function by using a machine learning algorithm, carrying out scoring screening on a compound to be screened and a compound butted with YTHDF1 molecules, carrying out affinity prediction, and screening out the small-molecule inhibitor targeting YTHDF1. According to the virtual screening method provided by the invention, the YTHDF1 targeting molecules with high affinity can be accurately screened on a large scale, candidate molecules are provided for research and development of medicines for treating YTHDF1 related diseases, and a foundation is laid for clinical tests and medicine optimization.

The invention belongs to the field of natural language processing technology and false news fact verification, and particularly relates to a novel false news detection method of a knowledge graph assistance and attention mechanism, which comprises the following steps: firstly, extracting data from a target news text, identifying a key entity and linking the key entity to a knowledge graph; obtaining an entity embedding vector based on the knowledge graph through a gating mechanism; then, taking sentences, topics and entities in the target text as input vectors, and extracting high-order knowledge representation of the input vectors by using a Transform encoder; then, obtaining deeper text embedding through a multi-head attention mechanism, including sentence-entity and sentence-theme embedding representation; and finally, comparing all the obtained advanced knowledge representations, and inputting the advanced knowledge representations into the model together to predict the truth score of the news, thereby effectively discriminating potential false news reports.

The invention provides an intelligent fire-fighting equipment fault identification method and system, and relates to the technical field of fire-fighting equipment fault identification, and the method comprises the steps of data collection, model construction, model training, fault identification and the like. The system comprises a data acquisition module, a model construction module, a model training module and a fault identification module, and through construction of a fire-fighting equipment fault identification model, interference of mechanical characteristic data of fire-fighting equipment in a vibration and impact environment state is overcome in a vibration or impact environment, a real equipment fault signal is identified, and the fire-fighting equipment fault identification module is used for identifying the fault of the fire-fighting equipment. The occurrence of fault misjudgment conditions is reduced, and the accuracy and reliability of fire fighting equipment fault recognition in the vibration impact environment state are improved.

The invention discloses an intelligent patent retrieval method and system based on a large language model reordering technology, objects processed by the method comprise a variety of input data including patent public numbers, free texts and text files, key information is extracted, query texts are constructed, retrieval recall is executed, sparse and dense vector retrieval technologies are combined, and the retrieval efficiency is improved. And results are fused through a min sum merge algorithm. The de-noising step uses a sliding window strategy and vector coding to extract information fragments related to the query text. And finally, by training a reordering model and finely adjusting a recommendation list, the ranking of retrieval results is optimized, and the accuracy and correlation of retrieval are improved. According to the intelligent retrieval method, the patent retrieval efficiency and accuracy are improved.

The invention belongs to the technical field of wind power generation. The invention provides a wind power plant road line selection method and system, a medium, a program product and equipment, and the method comprises the steps: inserting a plurality of vertexes between a starting point and a target point according to the map data of a target wind power plant to obtain a plurality of road sections; according to the road construction cost, the road longitudinal slope coefficient, the road longitudinal slope gradient, the road turning radius coefficient, the road earthwork cost coefficient, the filling amount, the excavation amount, the unit excavation cost and the unit filling cost of each road section, an objective function of the weighted road construction total cost is constructed; and performing optimization by minimizing the objective function to obtain a wind power plant road optimal line selection result. According to the method, under the constraint conditions of the maximum gradient, the minimum turning radius and the like, line selection arrangement is automatically carried out by taking the total construction cost of the wind power plant road as an optimization target, and the optimal scheme of the wind power plant road can be obtained more efficiently and more precisely.

The invention belongs to the field of rice genetic breeding, and discloses a breeding method of high-quality disease-resistant large-grain two-line sterile line rice, which ingeniously fuses two major technical fields of molecular biology and rice genetic breeding. A two-line sterile line A with an excellent genetic background is selected as a female parent and a restorer line B is selected as a male parent for hybridization, and then individuals carrying target disease-resistant and grain weight genes are accurately screened by using a molecular marker-assisted selection technology. By combining strict resistance inoculation test and appearance quality evaluation, the bred sterile line not only has excellent disease resistance (resistance to rice blast and bacterial blight), but also has large grain shape and high appearance quality. In addition, through multi-generation shuttling and generation-adding propagation, the hereditary character is further stabilized. The breeding method disclosed by the invention is efficient and accurate, provides a new thought and technical means for the field of rice genetic breeding, and is of great significance for improving the yield and quality of rice.

The invention provides a business demand generation method and device and electronic equipment, and relates to the technical field of data processing. The business demand generation method comprises the following steps: acquiring an acquisition data set of a business demand, and determining the use frequency of the business demand according to the acquisition data set; obtaining a service function to which the service demand belongs and grade information of the service function; determining key business demands according to the attribute information, the use frequency and the grade information, and determining target key business demands according to feature information of the key business demands; determining nodes to be optimized and optimization key information in a demand path according to the demand path of the target key business demand; the chapter information of the demand document is obtained, and the demand document of the target key business demand is generated in combination with the to-be-optimized node and the optimization key information, so that the actual demand of the user can be more accurately obtained, and the more accurate demand document is generated.

The invention discloses a machine learning-based isowurtzitane cage compound reaction substrate prediction method, which comprises the following steps: establishing a database containing condensation reactions of glyoxal and primary amine reaction substrates, marking the condensation reaction of which the product is an isowurtzitane cage compound as 1, and marking the condensation reaction of which the product is other types of compounds as 0; the method comprises the following steps: performing quantum chemistry calculation on primary amine reaction substrates, collecting descriptors describing chemical information, screening feature descriptors, and performing normalization processing; taking the feature descriptor as an input variable, taking a condensation reaction product marking result as an output variable, and selecting a random forest and a factorization machine to carry out weight distribution to construct an optimal fusion model; and performing prediction by using the optimal fusion model. According to the method, the reactivity of the primary amine reaction substrate can be quickly and accurately predicted, the research efficiency is improved, the experiment cost is reduced, and effective guidance is provided for synthesis of the isowurtzitane cage compound.

The invention belongs to the technical field of synthetic biological probiotic screening, and particularly relates to a synthetic biological probiotic screening method and system.The method comprises the steps that a target area in a DNA sequence is searched, K-mer frequencies in the target area are counted, the frequencies of K-mers with the same basic groups at the first k positions are sequentially put into a set according to the letter sequence of the residual basic groups of the K-mers, and the K-mers are screened; splicing all the sets into a frequency characteristic matrix of the target area in rows according to the alphabetic sequence of the first k bases, forming a multi-channel frequency characteristic matrix by the frequency characteristic matrixes of all the target areas, and inputting the multi-channel frequency characteristic matrix into a neural network model with a channel attention module for training; and obtaining a multi-channel frequency characteristic matrix of the to-be-screened synthetic biological probiotics from the DNA sequence of the to-be-screened synthetic biological probiotics, and inputting the multi-channel frequency characteristic matrix into the trained neural network model to obtain scores of the synthetic biological probiotics. The primary screening device can improve the primary screening efficiency of the synthetic biological probiotics.

The invention discloses a wood frog antibacterial peptide and application thereof in antibiosis, anti-inflammation and antioxidation, and belongs to the technical field of biology. The method comprises the following steps: firstly, analyzing skin mucus of wood frogs in Northeast by utilizing a peptidomics technology to obtain a large amount of peptide fragment information; and then, comparing the peptide fragment information with a known antibacterial peptide database, and screening out peptide fragments which possibly have antibacterial activity. Furthermore, through experiments, an antibacterial peptide which is extremely low in hemolytic activity and highly stable to pH and temperature and has anti-inflammatory and anti-oxidation effects is screened out. The amino acid sequence of the antibacterial peptide is as shown in SEQ ID NO: 3, and the antibacterial peptide lays a foundation for enriching the types of antibacterial peptides and adds new advantages for the application of the antibacterial peptide.

The invention relates to the technical field of geological exploration sampling analysis, in particular to a geological exploration sampling device and a sampling analysis method. The method comprises the steps of firstly obtaining geological exploration data of a target sampling position; further acquiring a local time domain window at a target acquisition moment; further acquiring a local speed change rate, a rock stratum number and a torque accumulation parameter in a local time domain window at each acquisition moment, and forming a feature vector of a target sampling position according to a time sequence of the sampling moments; further obtaining a data anomaly coefficient of the target sampling position according to the difference characteristics of the feature vector of the target sampling position and the feature vectors of other sampling positions in combination with the distance between the target sampling position and other sampling positions; and finally, screening out abnormal sampling positions according to outlier distribution characteristics of data abnormal coefficients of all sampling positions.

The invention relates to the technical field of nursing personnel distribution, in particular to a wireless call-for-help communication system for a smart ward. When a help calling signal is received, the system obtains all related data information including nursing experience, position information and the like of a nurse. Calculating a first matching factor (based on nursing experience) for each nurse and each patient by analyzing historical nursing data; meanwhile, a second matching factor (based on the position relation) is calculated according to the movement track of the nurse; the system also evaluates the accumulated workload and the future predicted workload of the nurse to reflect the busy degree and the fatigue degree of the nurse. In the help calling matching module, the nursing experience, the position relation, the workload and the task of the day of the nurse are comprehensively considered, and the target nurse most matched with the patient to be rescued is screened. According to the method, multiple factors are combined, more flexible and more accurate screening is achieved, it is ensured that the selected nurse can reach the position of the patient as soon as possible, the problem is effectively solved, and the efficiency and quality of medical care are improved.

The invention relates to the technical field of data processing, in particular to an industrial control protocol field content-based classification and grading method, which comprises the following steps of: acquiring original data streams of industrial equipment, performing primary data processing to obtain corresponding standard data streams, analyzing data frames in the standard data streams through a protocol analyzer, and storing the data frames in the standard data streams; the classification type of the standard data stream is determined according to the field content, key fields of all the data frames are extracted, feature vector groups are defined according to service requirements to determine key features of the data frames based on the key fields, and the data positioning trend is determined according to the classification type and the key features of the standard data stream to determine a grading method; and storing the classification type and the classification result in a database and displaying the classification type and the classification result through a visual interface. According to the invention, the communication protocol of a specific industry can be directly analyzed, and the data in the industrial control system can be identified and processed in a more targeted manner.

The invention discloses an order-driven mixed box stacking method, system and equipment and a medium. According to the method, a stacking process is decoupled into two associated subtasks of active warehouse-out decision and real-time stacking execution; the problems of frequent stack transfer and invalid reciprocating carrying caused by scattered goods combination in a physical layout limited scene in an existing offline stacking algorithm are effectively solved, and meanwhile, the space utilization rate bottleneck caused by lack of order global view in a traditional online stacking algorithm is also overcome; and a brand-new framework support and a global optimization decision view angle are provided for efficient disassembly operation in a limited physical space. An energy-based warehouse-out sequence optimization algorithm is introduced into a warehouse-out decision-making layer, and the algorithm deeply fuses the local fine sampling capability of Langevin dynamics and the global evolution capability of a genetic algorithm, so that the technical defects that a traditional heuristic rule is difficult to capture complex geometric features of goods and is easy to fall into local optimum are overcome; and the three-dimensional warehouse is guided to realize active and efficient replenishment scheduling.

The invention discloses an intracranial electrode lead track generation method, device and system, a medium and a product. The method comprises the following steps: acquiring medical image data of the brain of a patient after an electrode lead is implanted into the intracranial of the patient; extracting a feature point set corresponding to the electrode lead from the medical image data; according to trajectory features of the electrode lead in the cranium of the patient, a target feature point set implanted into the cranium of the patient is screened out from the feature point set, and the trajectory features at least represent that the bending degree of the electrode lead in the cranium of the patient does not exceed a preset bending value; and according to the target feature point set, simulating to obtain an intracranial electrode lead track of the patient. According to the method, the intracranial electrode lead wire part can be accurately identified by utilizing the intracranial trend of the electrode lead wire, and accurate three-dimensional visual rendering is performed on the intracranial electrode lead wire through the point corresponding to the intracranial electrode lead wire part, so that a doctor can better understand the position relationship between the intracranial electrode lead wire and the surrounding nerve fiber bundle, and the accuracy and the reliability of the system are improved. Doctors can optimize program control parameters, and the treatment effect is improved.

The invention relates to the technical field of data processing, in particular to a blowing flow adjusting system in the metal powder machining process. The system comprises the steps of collecting flow data, obtaining a mutation factor of each data in a main body flow data sequence, and obtaining a noise factor of each data in the main body flow data sequence according to the mutation factor of each data in the main body flow data sequence; based on the noise factor, adaptively obtaining a judgment threshold value of a Mahalanobis distance of each datum in the main body flow data sequence; according to the judgment threshold value of the Mahalanobis distance of each data in the main body flow data sequence, noise data is obtained for smoothing, the denoising effect is improved, and the influence of abnormal flow data on noise data recognition is avoided.

This invention relates to the field of medical information management technology, including an AI-based endometriosis management system. The system comprises a thickness dynamic monitoring module, an inflammation level assessment module, a lesion spread analysis module, a periodic lesion tracking module, and a disease progression analysis module. In this invention, AI is used to segment ultrasound image data, enabling dynamic monitoring of lesion areas. Lesion boundaries are extracted, and changes in local inflammatory marker concentrations are calculated, making the screening of abnormal inflammatory areas more accurate and enabling early identification of lesion development. The tracking of periodic lesions, combined with analysis of lesion contour changes and area increases / decreases, enhances the quantitative control of disease progression. The future development trend of lesions, combined with the calculation of periodic inflammatory pathways, makes the prediction of disease evolution more consistent with physiological changes. Multi-dimensional lesion analysis combined with AI intelligent calculation expands the diagnosis and treatment of endometriosis from static assessment to dynamic trend prediction, improving the adaptability of personalized treatment plans.

The present application relates to the technical field of fault detection, in particular to a chemical production process abnormal condition identification method and system, comprising the following steps: collecting reactor temperature concentration and flow and aligning and denoising to generate a multi-source time sequence set, extracting spatial measurement point temperature difference and differentiating symbols to generate a temperature difference direction sequence, calculating conversion amount based on concentration and flow and calculating rate gradient and direction sequence inversion comparison to generate heat release inversion sequence, obtaining inversion time and spatial coordinates and misaligning to align to form a lag mismatch set, combining time window temperature change and mismatch frequency to input a support vector machine classification to obtain abnormal condition identification results, in the present application, a multi-source synchronous time sequence set is constructed to extract temperature difference direction, spatial gradient is calculated to obtain inversion characteristics, time and space misalignment is implemented to align to capture physical conduction lag law, mismatch frequency is used to rely on a support vector machine to perform classification, single point limitations in dealing with conduction lag are overcome, the technical problem that hidden faults are easily covered is solved, and accurate identification of abnormal variations in complex conditions is realized.

The invention belongs to the field of genetic breeding, and relates to cultivation of powdery mildew resistant wheat. The invention provides a detectable molecular marker for tracking a powdery mildew-resistant gene PmDR144 and application of the detectable molecular marker in disease-resistant breeding. The molecular marker comprises an upstream primer HENU624-F and a downstream primer HENU624-R, the nucleotide sequence of the upstream primer HENU624-F is as shown in SEQ ID No. 1, and the nucleotide sequence of the downstream primer HENU624-R is as shown in SEQ ID No. 2. The primer provided by the invention can be used for detecting whether a wheat breeding material carries the powdery mildew resistant gene PmDR144 or not before the powdery mildew in the wheat seedling stage is attacked, a variety carrying the powdery mildew resistant gene PmDR144 is purposefully selected, and the marker not only is rapid and accurate in detection, but also saves the cost, improves the breeding efficiency and accelerates the breeding process.

The invention discloses a method for breeding a new red-pulp seedless grape variety based on molecular markers, and belongs to the field of biological breeding. The molecular marker obviously related to red pulp and seedless characters of grapes is excavated through a whole genome analysis method, the molecular marker has broad spectrum, and red pulp and seedless character phenotypes of grapes can be accurately identified in the early stage of hybrid seedlings by amplifying electrophoretic band types of gene segments of the molecular marker, so that the method is simple and convenient to operate. Compared with traditional breeding and cultivation breeding, the method saves at least five years of breeding period, obviously shortens breeding period, saves a large amount of land and management investment, improves breeding efficiency, and provides support for efficient molecular assisted breeding of new red-pulp seedless grape varieties.

This invention relates to the field of nuclear radiation detection technology, and more particularly to an energy spectrum processing method for a quasi-hemispherical CdZnTe detector. The method includes: acquiring the raw signal collected by the quasi-hemispherical CdZnTe detector; normalizing the amplitude of the raw signal from the baseline to the highest point during the rise phase; dynamically selecting multiple energy thresholds; performing rise time discrimination on the normalized signal within each energy threshold; filtering out signals with rise times greater than RTflag and determining the rise time range corresponding to that energy threshold; constructing an intermediate energy spectrum corresponding to each energy threshold based on the signals with rise times greater than RTflag and the rise time range; and post-processing the intermediate energy spectrum to obtain the target energy spectrum corresponding to each energy threshold. This invention combines energy threshold discrimination, rise time discrimination, and energy spectrum post-processing algorithms to stably improve the energy resolution of the quasi-hemispherical CdZnTe detector for 662 keV gamma rays from the conventional 2%~3% to less than 1%.

The invention belongs to the technical field of drug detection, and particularly relates to a method for detecting lycium barbarum glycopeptide metabolites in rat plasma and tissues based on non-targeted metabonomics. According to the method, after rat lycium barbarum glycopeptides are administered through intragastric administration, blood, liver, intestine, lung and brain tissue samples after administration are collected. A UPLC-Q-TOF-MS (Ultra Performance Liquid Chromatography-Quadrupole Time of Flight Mass Spectrometry) analysis method is combined with a statistical analysis method comprising principal component analysis (PCA) and orthogonal-partial least squares discriminant analysis (OPLS-DA) to carry out in-vivo metabonomics analysis on the wolfberry glycopeptide. Then, based on a human metabolome database (HMDB), lycium barbarum glycopeptide differential metabolite is searched, so that the pharmacological action mechanism of the lycium barbarum glycopeptide is clarified from the perspective of metabolism, and a foundation is laid for deep research and development of the medicinal value of the lycium barbarum glycopeptide in future.

The present invention relates to the field of data security analysis technology, and specifically, to a data security analysis system and method based on artificial intelligence. It includes: a data processing unit performs parallel cleaning and standardization processing on network traffic data, user behavior data, and system log data, and integrates them into a comprehensive feature vector; an anomaly detection unit uses the DBSCAN algorithm to perform cluster analysis on the comprehensive feature vector, and uses the Gaussian mixture model to perform probability modeling to screen outliers; a behavior analysis unit determines the type of outliers through pattern matching, and performs risk assessment on the outliers to generate a security analysis report; a security response unit adopts an automated response mechanism based on the risk level of the outliers, and generates a security response report. The data security analysis system and method based on artificial intelligence achieves efficient outlier division by dynamically adjusting clustering parameters and density sensing technology, and then combines the Gaussian mixture model with the screening mechanism of posterior probability to achieve anomaly detection and security risk assessment.