37 results about "Confidence set" patented technology

The technology disclosed relates to detection of anonymous proxies and bots making requests to a cloud based resource on the Internet, such as a web server or an App server. The technology can leverage one or more of: instrumentation of web pages that samples response times and other characteristics of communications by a requestor device over multiple network segments; lack of prior appearance of the requestor device across multiple, independently operated commercial web sites; and resolver usage by the requestor. These signals can be analyzed to score a requesting device's reputation. A location reported by a user device can be compared to a network characteristic determined location.

The invention discloses a dispatching method for achieving robust operation of an electrical power system. The dispatching method comprises the steps that S1, original data information is obtained; S2, under a certain confidence coefficient level, an upper limit and a lower limit of a mean value of day-ahead, intra-day and real-time wind power generation forecast errors, an upper limit and a lower limit of day-ahead, intra-day and real-time photovoltaic power generation forecast errors, and an upper limit and a lower limit of day-ahead, intra-day and real-time load forecast errors are obtained; S3, a day-ahead dispatching plan, a robust safe operation range corresponding to the day-ahead dispatching plan, an intra-day dispatching plan, a robust safe operation range corresponding to the intra-day dispatching plan, a real-time dispatching plan and a robust safe operation range corresponding to the real-time dispatching plan are obtained. According to the method, the rolling coordination technologies of forecast information, current operation information and historical operation information are considered simultaneously, the robust safe operation ranges of the system are obtained, and therefore the dispatching plans are not limited to a unique preset value, and flexible dispatching in the robust ranges can be achieved. The obtained dispatching plans can be used for coping with stochastic volatility of new energy power generation better, and safety and economical efficiency are both considered.

The embodiment of the invention discloses an anti-disturbance generation method and device for an object detection model. The method comprises the following steps: acquiring a first anti-disturbance and a first training sample set; And determining a first confrontation sample according to the first training sample and the first confrontation disturbance in the first training sample set, and performing first confrontation disturbance correction on the first confrontation disturbance based on a first target object confidence set corresponding to the first confrontation sample determined by the object detection model to obtain a second confrontation disturbance. After a new training sample is obtained from the first training sample set each time, new anti-disturbance is obtained through correction again based on the new training sample and the anti-disturbance obtained through last anti-disturbance correction. And when the N confrontation disturbances obtained by N times of correction areconverged, determining the (N + 1) th confrontation disturbance obtained by the Nth confrontation disturbance correction as the target confrontation disturbance corresponding to the object detectionmodel. By adopting the embodiment of the invention, the efficiency and the applicability of the anti-disturbance generation method can be improved.

The invention discloses a hyperspectral image semi-supervised classification method based on comprehensive confidence. The method comprises the following steps: reading a hyperspectral image; Calculating a graph weight matrix; 8, performing adjacent connection on the sparse graph weight matrix; Calculating a normalized graph weight matrix; Obtaining an initial training set and a candidate set; Setting collaborative training iteration times and starting a training process; Training a polynomial logic regression classifier; Obtaining a prediction label of the candidate set sample by using a polynomial logic regression classifier; Obtaining prediction tags of the candidate set samples by using a semi-supervised graph classification method; Selecting two candidate samples with consistent prediction tags and corresponding prediction tags to form a protocol set, and forming a comprehensive confidence set by corresponding confidence coefficients; Screening out a protocol set sample with a comprehensive confidence coefficient higher than 99% and a corresponding prediction label, and forming an amplification set and adding the amplification set into a training set; Removing an amplificationset sample in the candidate set; And judging whether the training reaches a set number of times, if not, continuing iteration, and if yes, stopping iteration, and classifying the hyperspectral imagesby using the semi-supervised graph.

A text mining apparatus, a text mining method, and a program are provided that accurately discriminate inherent portions of each of a plurality of text data pieces including a text data piece generated by computer processing.

A text mining apparatus 1 to be used performs text mining using, as targets, a plurality of text data pieces including a text data piece generated by computer processing. Confidence is set for each of the text data pieces. The text mining apparatus 1 includes an inherent portion extraction unit 6 that extracts an inherent portion of each text data piece relative to another of the text data pieces, using the confidence set for each of the text data pieces.

The invention provides an acquisition method and system of a sample picture data set. The method comprises the following steps: executing the cleaning process of picture data to be cleaned: acquiringa positive sample picture set and a negative sample picture set, wherein the feature information of picture data in the positive sample picture set is the same as that of a target picture, and the feature information of picture data in the negative sample picture set is different from of the target picture; performing training according to the positive and negative sample image sets to obtain a neural network sorter; classifying the picture data to be cleaned according to the neural network sorter to obtain a plurality of confidence sets; and executing an acquisition process of the sample picture data set: acquiring picture data in the confidence sets of which the confidence levels reach a preset level to obtain the sample picture data set. Through adoption of the acquisition method and system, automatic screening and classification are realized to obtain the sample picture data set, and the screening and classification efficiency and accuracy are increased.

The invention discloses an online multi-target tracking method based on track metric learning. The method is used for solving the technical problem that an existing online multi-target tracking methodis poor in practicability. According to the technical scheme, the method comprises: firstly, an existing target detection algorithm is used for generating a detection response; and then, dividing theexisting track set into a high-confidence set and a low-confidence set, processing a data connection problem of the high-confidence set and the next moment detection response by using static featuresand a traditional measurement method, and enhancing a data connection capability for the low-confidence set by using a similarity measurement matrix to obtain a final result. Existing track information is taken as a training sample set, a similarity measurement matrix between tracks and detection responses is learned online, and the resolution capability of track discrimination is enhanced. The technical problems that accurate data connection is difficult to perform and the tracking effect is seriously limited by the detection effect in the background technology are solved, the multi-target tracking effect is improved, and the practicability is good.

A method for performing advanced driver assistance system (ADAS) functions for a vehicle includes receiving a plurality of inputs from a plurality of sensors disposed on the vehicle, determining a pool of location candidates of the vehicle based on a first subset of the plurality of inputs, iteratively updating the pool of location candidates based on a second subset of the plurality of inputs, generating an estimate of vehicle location based on an average of location candidates, evaluating a confidence level of the estimate of vehicle location, generating a beyond line-of-sight situation awareness based on the plurality of inputs and locating the vehicle on a digital map based on the confidence level and the estimate of vehicle location, and performing ADAS functions based on the beyond line-of-sight situation awareness and location of the vehicle on the digital map.

The embodiment of the invention provides an image processing method. The method comprises the steps of extracting image features of a to-be-processed image to obtain a first feature value matrix; processing the first characteristic value matrix by adopting a classification prediction model, determining a prediction confidence coefficient of the to-be-processed image relative to each predeterminedangle category in a plurality of predetermined angle categories, and generating a prediction confidence coefficient set, the predetermined angle category indicating an angle interval where an offset angle is located; determining an offset angle of the to-be-processed image according to the prediction confidence set; and rotating the to-be-processed image according to the offset angle. According tothe method, the problem of determining the deviation angle of the image is converted into the angle classification task, so that the calculation complexity can be effectively reduced, and the accuracy of determining the deviation angle is improved. In addition, the embodiment of the invention further provides an image processing device, a medium and computing equipment.

The invention discloses a multi-target tracking algorithm based on tracklet confidence and appearance learning. The multi-target tracking algorithm includes: firstly, reading in an image and detectioninformation thereof, calculating the similarity of a detection result and obtained tracklets, performing local association computation on the tracklets and the detection result, then calculating thesimilarity among all the tracklets, and performing global association computation among the tracklets so that the robust and rapid online multi-target tracking algorithm is realized. According to themethod, the targets can be accurately and rapidly distinguished, an appearance model is updated by employing tracking results in an incremental manner, and the tracklets can be successfully associatedin the shielding condition.

The invention discloses an optical module fault diagnosis and early warning method, a device and a system which are used for solving the technical problems that in the prior art, the fault probabilityof an optical module cannot be given, and the adaptability of a model for judging whether the optical module breaks down through a static threshold value method is limited. The method comprises the following steps: acquiring digital diagnosis detection DDM data of a first period of the optical module in real time, processing the DDM data of the first period by adopting an anomaly detection model,and determining a first anomaly vector; according to the first exception vector, calling a sample frequent item set of an exception detection model and a corresponding sample confidence set, and judging whether the first exception vector is matched with an exception vector in the sample frequent item set; if yes, determining that the fault probability of the optical module in the first period isthe sample confidence corresponding to the abnormal vector in the sample frequent item set; and if not, determining that the optical module has no fault in the first period.

The specification provides a commodity identification and settlement method, device and system. The method comprises the following steps: acquiring product images collected by a camera, and dividing the product images into image regions corresponding to the respective products; performing image recognition on the image regions to obtain a confidence set corresponding to each of the image regions,and determining whether the image areas identify barcodes, wherein the confidence set is used to describe the confidence of the at least one product identifier corresponding to the image regions; performing text recognition on the image regions to obtain text information of the image regions; and determining, according to a preset policy, product identifiers corresponding to the image regions, wherein the preset policy is set based on the confidence, whether barcodes are associated, and the text information. The application corrects the confidence set by collecting multi-dimensional data information, which increases the recognition efficiency and recognition accuracy, and improves the practicality and user experience.

A text mining apparatus, a text mining method, and a program are provided that enable the influence that computer processing errors have on mining results to be reduced during text mining performed on a plurality of text data pieces including a text data piece generated by computer processing. A text mining apparatus 1 to be used includes an inherent portion extraction unit 6 that, for each of a plurality of text data pieces including a text data piece generated by computer processing, extracts an inherent portion of the text data piece relative to another of the text data pieces, an inherent confidence setting unit 7 that, for each inherent portion of each of the text data pieces, sets inherent confidence indicating confidence of the inherent portion, using the confidence that has been set for each of the text data pieces, and a mining processing unit 8 that performs text mining on each inherent portion of each of the text data pieces, using the inherent confidence.

The invention discloses a confidence-based voucher matching method and system. The method comprises the steps of 1, obtaining contract elements of a to-be-matched contract; 2, matching the contract tobe matched with candidate invoices according to contract elements, constructing a basic confidence set based on a matching result, and determining confidence C1 of each candidate invoice If in the basic confidence set; And step 3, sorting all the candidate invoices If in the basic confidence set according to the confidence C1. According to the voucher matching method, the contract to be matched is matched with the candidate invoice, and the matching results are sorted according to the confidence degree, so that the reliability and efficiency of voucher matching can be improved.

The invention discloses an online cooperative positioning method and system based on Wi-Fi RSS, and belongs to the technical field of communication and wireless networks. The method is characterized by using the RSS of other users requesting positioning at the same time to obtain the cooperative confidence coefficients of all candidate reference points of a target user, so as to enrich positioninginformation without adding extra hardware equipment, so that the positioning precision is improved; integrating the signal distance and the cooperative confidence coefficient of the user, determininga final positioning result of the target user according to the obtained final comprehensive confidence coefficient of the candidate reference points, and improving the positioning reliability and accuracy; eliminating signal sources with low regional acceptance rates or weak overall signal intensity, so that remote signal sources and temporary hotspot sources are eliminated and the positioning precision is improved, wherein such invalid signal sources often bring interference to fingerprint comparison work; and standardizing the signal distance set and the collaborative confidence set of thesystem to eliminate the dimensional influence of the two indexes, so that the positioning accuracy is improved.

The invention discloses an information distribution strategy optimization method based on a dynamic feedback mechanism in Internet of Vehicles. The method comprises the following steps that S1, a system initializes an attribute set, a confidence set of a strategy maker group and a background probability vector of each attribute; S2, a strategy maker group generates an initial strategy, after the vehicle receives the message, whether the message is valid or not is fed back, and a road side unit collects attribute information and feedback data of the vehicle; S3, the road side unit deletes the abnormal feedback data; S4, an optimal strategy of the information distribution scene is determined according to a decision tree algorithm; S5, the road side unit updates the confidence set of the strategy maker group and the background probability vector of each attribute according to the optimal strategy; S6, the road side unit obtains the optimized information distribution strategy, encrypts themessage according to the optimized information distribution strategy, and broadcasts the encrypted ciphertext information to the vehicle. According to the optimization method provided by the invention, the information distribution strategy is effectively optimized by designing a dynamic feedback mechanism, and the information distribution accuracy is improved.

The invention discloses a voice verification method and device, electronic equipment and a computer readable storage medium. The method comprises the steps of collecting a sound signal in real time; extracting audio features of the sound signals; inputting the audio features into a multi-classification verification model, and outputting a confidence coefficient set through the multi-classificationverification model, the confidence coefficient set comprising confidence coefficients of non-command words and confidence coefficients of a plurality of command words; selecting the maximum confidence in the confidence set; based on the condition that the maximum confidence coefficient is not the confidence coefficient of the non-command word, loading a binary classification verification model corresponding to the target command word to which the maximum confidence coefficient belongs; inputting the audio features into a binary classification verification model, and outputting a target confidence coefficient of the target command word through the binary classification verification model; and executing a function matched with the target command word based on the condition that the target confidence coefficient is greater than a specified threshold. Through the technical scheme of the invention, the equipment can ensure the accuracy of command word detection while saving power consumption and system resource consumption.

The invention relates to an artificial intelligence technology, and discloses a face selection method, which comprises the following steps: performing face confidence calculation processing on each face in a scene picture to obtain a face confidence; selecting human faces with human face confidence higher than a preset human face confidence threshold from the scene picture to obtain a first human face set and a human face confidence set of the first human face set; determining a self-adaptive confidence coefficient threshold value through a real face confidence coefficient floating range; selecting human faces with human face confidence higher than an adaptive confidence threshold from the first human face set as a second human face set; and selecting the face occupying the largest area in the scene picture from the second face set as a face selection result of the face recognition object. The invention also relates to a block chain technology, and the preset face confidence threshold value is stored in the block chain. The problems that in the prior art, the accuracy of a face selection result used for face recognition is poor, and the possibility of deleting a real result by mistake exists can be solved.

The invention provides an attribute-based multi-party strategy fusion method in an Internet of Vehicles information distribution scene, and solves the problem that a traffic information service cannoteffectively formulate an appropriate distribution strategy, and the method comprises the following steps: S1, initializing an attribute set and a confidence set of a strategy formulator group by a system; S2, enabling the road side unit to preprocess the strategies formulated by the strategy formulator group and unify strategy formats; S3, enabling the road side unit to screen out conflict sub-policies and non-conflict sub-policies by matching the policies, and unifying the number of attributes in the conflict sub-policies; S4, enabling the road side unit to perform multi-party strategy fusion according to the confidence coefficient of the strategy maker group and the conflict attribute values to form a fusion sub-strategy; S5, enabling the road side unit to merge the non-conflict sub-strategy and the fusion sub-strategy to form a fused access control strategy, performing message encryption according to the access control strategy, and broadcasting encrypted ciphertext information toa message receiver. The method provided by the invention effectively improves the accuracy of information broadcast distribution.