44results about How to "Full excavation" patented technology

The invention discloses a group user mining method and apparatus. The method comprises the steps of forming daily stay points of users by daily track points, meeting preset time threshold, distance threshold and direction change threshold conditions, of the users according to historical track data, consisting of the track points, of the users, and classifying the daily stay points of the users into multiple public stay regions with semantic information; based on the stay points meeting the preset time threshold, distance threshold and direction change threshold conditions and the public stay regions with the semantic information, performing statistics on most frequent travel starting time period track feature attributes, most frequent travel ending time period track feature attributes, most frequent stay region list track feature attributes and most frequent repeated track mode track feature attributes of the users; and clustering the four track feature attributes of the users into multiple group users by adopting a preset clustering method. Therefore, the group users can be mined more comprehensively based on the historical track data.

The invention discloses a vulnerability mining detection method, server, device and system. The method comprises the following steps: acquiring suspicious sample logs uploaded by clients; performing duplication elimination on the acquired suspicious sample logs; and inputting suspicious sample logs left after duplication elimination into a sandbox, performing scanning to obtain a scanning log, and determining whether samples corresponding to the suspicious sample logs have vulnerabilities or not according to the scanning log. In the scheme, the clients are taken as probes spreading all over the Internet to collect the suspicious sample logs corresponding to suspicious files, so that relevant information of the samples in which vulnerabilities exist possibly in the Internet can be mined comprehensively. Then, samples in which security vulnerabilities exist really are determined from the samples in which the vulnerabilities exist possibly according to the scanning log, so that actual security situations of the samples are determined. Scanning is performed in the sandbox on a server side, so that an enclosed running environment is provided for the suspicious sample logs, and the security of the server side in a detection process is ensured.

The invention discloses a sensor data classification method based on phase space reconstruction, and the method comprises the steps: 1), mapping a one-dimensional sensor time sequence to an m-dimensional phase space through the phase space reconstruction based on coordinate delay so that sensor data exposes implicit key information to acquire a phase space track matrix; 2) inputting the extracteddata of each sensor at the moment i into a long-short-term memory network module, and setting the time step length input by the long-short-term memory network to be the embedded dimension m of phase space reconstruction; and 3) inputting the RGB matrix of the state information analogy image output by each long-term and short-term memory network into a two-dimensional convolutional neural network module, and inputting features extracted by the two-dimensional convolutional neural network into a full connection layer to obtain a classification result. The problems of limitation of an original one-dimensional sensor data structure and limitation of sensor data classification accuracy are solved, and the sensor data at the time point can be accurately classified.

The invention discloses a social recommendation method based on a multi-feature heterogeneous graph neural network. The method comprises the following steps: extracting various attribute information of social network users and topics for coding; processing the user coding information and the topic coding information through a multi-layer perceptron to obtain initial feature vector representation of each user and topic; establishing a heterogeneous graph by taking users and topics as nodes, inputting the heterogeneous graph into a heterogeneous graph neural network, performing information transmission in the graph in combination with an attention mechanism, and updating feature vector representation; and performing similarity calculation on the user feature vectors, and selecting top-k users and top-k topics with the highest similarity with the user vectors for recommendation. Various types of attribute information of the users and the topics is extracted at the same time, the users and the topics serve as nodes at the same time to establish the heterogeneous graph, social information can be mined more comprehensively, information transmission and aggregation are carried out through the heterogeneous graph neural network, features of the users and interested topics of the users are deeply fused, and recommendation accuracy and user experience are improved.

The invention discloses a fault monitoring method based on a sequence correlation locality preserving projection algorithm, and aims to infer a brand-new data feature mining algorithm and implement fault monitoring based on the algorithm. Specifically, the method comprises the following steps: combining time sequence correlation maximization and local neighbor structure minimization into a targetfunction; in the solving process, guaranteeing the mutual orthogonal characteristic between the projection transformation vectors; and finally, executing fault monitoring by utilizing the extracted potential features and model errors. Compared with a conventional method, a sequence correlation locality preserving projection algorithm involved in the method is a brand-new feature extraction algorithm, considers the autocorrelation features and local neighbor features at the same time in the projection transformation process, guarantees the orthogonal characteristic of projection transformationvectors, and can mine the hidden useful information in the training data more comprehensively. Therefore, the method provided by the invention is a more preferable fault monitoring method.

The invention provides a security vulnerability assessment method for a cigarette industry control system. The method comprises the steps of identifying data streams, and tagging data stream paths and stream directions of the data streams; classifying the data streams according to attributes of the data streams; analyzing an attack surface; judging whether the data streams have technical control measures corresponding to the data streams in the cigarette industry control system or not; if yes, judging whether equipment nodes on the data stream paths have security management measures corresponding to the equipment nodes or not; if yes, performing security vulnerability testing and analysis on key nodes on the data stream paths, performing statistics on security vulnerability points on the key nodes on the data stream paths, and compiling the security vulnerability points existent in the cigarette industry control system in a predetermined document format; and if the equipment nodes on the data stream paths do not have the security management measures corresponding to the equipment nodes, configuring the security management measures corresponding to the equipment nodes for the equipment nodes on the data stream paths, and if the data streams do not have the technical control measures corresponding to the data streams in the cigarette industry control system, configuring the technical control measures corresponding to the data streams for the data streams. According to the security vulnerability assessment method, the security vulnerability of the cigarette industry control system can be assessed more comprehensively and more accurately.

The invention discloses a GC-IMS pear production place distinguishing method based on an aromatic substance fingerprint spectrum. According to the invention, a headspace gas chromatography-ion mobility spectrometry instrument is used for detecting pears from different producing areas; GC-IMS sample data are obtained; the GC-IMS sample data is subjected to statistical analysis through LAV analysissoftware and GC * IMS Library Search Software qualitative software of the instrument, and GC-IMS fingerprint spectrums of volatile aroma components of the pears from all production places are obtainedfor distinguishing the production places of the pears. The GC-IMS technology is applied to collection and trace analysis of complex aroma substances in the pears from different producing areas, comprehensive fingerprint information is provided for distinguishing the producing areas of the pears, rapid and accurate distinguishing of the producing areas of the pears from different producing areas is achieved, and theoretical basis and data support are provided for identifying and tracing the producing areas of the pears.

The invention provides a physiological parameter measuring method based on a multi-scale fusion network. The method comprises the following steps of sampling physiological signals and generating a one-dimensional physiological signal data sequence; performing noise filtering on data fragments meeting the signal quality requirement through data fragment division and signal quality evaluation; performing mathematical transformation on the data fragments to generate a multi-dimensional input data tensor; extracting potential features from input data by utilizing the multi-scale fusion network to obtain estimated values of physiological parameters; by identifying a measurement mode identifier, taking a mean value of all the estimated values as a physiological parameter measurement value in a static mode; and taking a one-dimensional continuous data sequence formed by all the estimated values as a physiological parameter continuous measurement value in a dynamic mode. According to the method, complementary information of different scales in the signals can be fully extracted, accurate measurement of the physiological parameters is achieved, the application range of the method covers the measurement of all the physiological parameters, and the method has certain application value in the fields of cardiovascular disease research and signal processing research.

The invention discloses a combined cutterhead of a rectangular top pipe heading machine. The combined cutterhead comprises a rectangular cylinder. The outer edge of an excavation surface of the rectangular cylinder is provided with a peripheral guardrail which is arranged in a rectangular manner. Peripheral cutters are equidistantly welded at the front edge position of the peripheral guardrail. One circle of the excavation surface of the rectangular cylinder is provided with reinforcing blocks which are equidistantly welded on one circle of the excavation surface of the rectangular cylinder and are integrally formed with the inner wall of the peripheral guardrail. The middle part of the excavation surface and four hanging corners of the rectangular cylinder are respectively provided with amain installing hole and auxiliary installing holes. Through the combined design of the main cutterhead assembly and the auxiliary cutterhead assembly, an excavation area and excavation strength canbe enlarged; a main cutterhead assembly and an auxiliary cutterhead assembly have the same structure and respectively comprise a wheel hub and outer protecting ring; the axial position of the wheel hub is provided with a fish tail cutterhead; and through the fish tail cutterhead which projects in an axial direction, preliminary fragmentation excavation in a tunneling process of a pipe pushing jackcan be realized; and a pressure which is generated in a tunneling process can be shared through arranging a plurality of web plates.

The invention discloses an intrusion detection method based on a traceability graph, and belongs to the field of computer system security, and the method comprises the steps: obtaining a traceability graph of a to-be-detected behavior, and calculating the node importance; rapid judgment and accurate judgment are carried out. In the rapid judgment stage, N1 nodes with high importance and corresponding K1 neighborhood nodes are selected, the traceability graph is converted into a first neighborhood matrix through a mapping rule, so that main body features of the traceability graph are rapidly extracted, if the difference m1 between the features and normal rule behaviors is larger than a threshold ThH during detection, intrusion behaviors are judged, and if the difference m1lt is larger than a threshold ThH, the traceability graph is judged to be intrusive behaviors; if so, judging that the behavior is normal; if yes, judging whether TLt; tilt; Tilt; if yes, entering a precise judgment stage: expanding a central node N2 (N2gt; n1) and a neighborhood node K2 (K2gt; (K1) scale, deeply mining the traceability graph, constructing a second neighborhood matrix and extracting features, if the difference between the features and normal rule behaviors is m < 2lt >; if Th, determining the behavior as a normal behavior, otherwise, determining the behavior as an abnormal behavior. According to the invention, intrusion detection can be efficiently and accurately realized.

The invention relates to a human body posture estimation method based on a joint relation, and the method comprises the following steps: S1, constructing a joint relation module which comprises two sub-modules: a channel-based feature relation module and an adjacent joint space relation module; s2, constructing a human body posture estimation model based on a joint relation; s3, training a human body posture estimation model based on a joint relation by utilizing the marked human body posture data; and S4, performing a human body posture estimation task based on a single image by using the trained human body posture estimation model added with the joint relation module to obtain a predicted human body posture. Compared with the prior art, the invention effectively overcomes the problem that the positions of four-limb joints with high degree of freedom, such as wrist joints, ankle joints and shielded invisible joints, are difficult to detect in the image, and is high in human body posture estimation accuracy.

The invention provides a social group discovery method, a social group discovery device and a storage medium, which are used for improving the social group discovery precision and accuracy. The socialgroup discovery method comprises the following steps: searching candidate users from a communication data source by using a preset seed user and a constraint condition, the candidate users comprisinga first user communicating with the seed user and/or a second user satisfying the constraint condition; extracting associated feature information associated with the seed user from the communicationdata of the candidate user; determining confidence corresponding to each candidate user according to the extracted associated feature information; determining the candidate users of which the confidence coefficients are greater than a preset threshold value as target users; establishing an initial group network according to the communication relationship between the target users; and discovering asocial group sub-network from the initial group network by utilizing a group discovering algorithm.