161 results about "Prediction score" patented technology

The invention discloses a method for establishing a rough-arrangement search model based on LTR, and the method comprises the following steps: S1, collecting candidate total data, and completing the preprocessing operation; s2, the sorting score of each piece of data is calculated and arranged in an ascending order, and the data are divided into equivalent data buckets; s3, randomly extracting data to form a training sample pair, generating a Pair-wise data set, and marking a sorting relationship; s4, executing rough-arrangement search model training and calculating a prediction score; s5, constructing a loss function by adopting a Lambda gradient signal, feeding back a sorting error and iteratively updating model parameters; s6, monitoring a training performance index, stopping training after a stopping condition is met, and deploying the model to an online environment; and S7, storing the trained model file, and defining an input format, a sorting structure and an output interface. According to the method, the sorting accuracy and fine sorting consistency of the rough sorting search model are improved, and the generalization ability of the model is remarkably enhanced.

The invention discloses a sports market demand prediction method and device based on deep learning and a medium, and relates to the technical field of market demand prediction, and the method comprises the steps: collecting sports demand data, and carrying out the preprocessing; performing relation mining and graph structure learning on the preprocessed sports demand data through a graph attention space-time network to generate a macroscopic demand potential energy graph; performing potential area identification on the macroscopic demand potential energy diagram by adopting a pre-trained sports market simulation model, and outputting local demand prediction data; performing weighted fusion and error correction on the macroscopic demand potential energy map and the local demand prediction data, and outputting a sports demand prediction score; and making a sports market demand strategy according to the sports demand prediction score and the multi-granularity demand prediction report, and transmitting the sports market demand strategy to an enterprise manager through an enterprise decision support interface. According to the method, multi-level accurate prediction and decision support of sports market demands are realized through dual-mechanism cooperation of the graph attention space-time network and the space-time convolution.

The invention discloses a self-updating method and device of a financial risk control model, a storage medium and a terminal, relates to the technical field of data processing, can be applied to the field of financial risk control, and mainly aims at solving the problem that the model recognition accuracy is low due to the fact that an existing financial risk control model is not timely updated. The method mainly comprises the following steps: acquiring anomaly prediction scores and confidence coefficients of different newly-added samples obtained in a process of performing anomaly identification on newly-added sample data in a production environment by a financial risk control model; extracting a low-confidence sample from the newly added samples according to the abnormal prediction score and the confidence; performing clustering processing on the low-confidence samples, and constructing an updated training sample set according to target samples extracted from each cluster; and performing incremental learning update training on the financial risk control model based on the training sample set, so as to continue to execute anomaly recognition of subsequent sample data based on the financial risk control model completing update training. The method is mainly used for updating the financial risk control model so as to improve the timeliness of model updating.

The invention discloses a synthetic lethal prediction method based on non-generality forgetting and low-rank interaction multi-mode adaptive fusion, and relates to the technical field of bioinformatics, computational biology and drug target screening. The method comprises the following steps: firstly, constructing a multi-view gene embedding expression based on a gene graph structure, a hypergraph structure and a knowledge graph; then constructing a difference vector and an interaction vector for any gene pair so as to characterize the difference and potential complementary relationship between the genes; multi-source information fusion is realized by using a deep learning model containing interactive attention, a forgetting gate and a multi-modal fusion mechanism; on this basis, multi-task optimization is carried out through joint classification loss, modal interaction loss and redundancy suppression loss, and optimal discrimination is realized through a dynamic threshold search strategy; and finally, outputting a sorting result of the candidate synthetic lethal gene pairs according to the prediction score. According to the method, the accuracy and generalization ability of synthetic lethal relationship prediction can be effectively improved, and a high-reliability calculation auxiliary tool is provided for anti-cancer drug target screening and gene therapy strategy design.

The invention relates to the technical field of rehabilitation medical treatment, and discloses a multi-modal data analysis method for rehabilitation decision support. The method comprises the steps that current physiological index data of a patient are collected and compared with historical samples, and individual difference evaluation values are obtained; analyzing current physiological index data and training intensity related branch paths through a tree decision model, and determining a risk prediction score; action priorities are calculated in combination with the risk scores and the physiological data, and a preliminary action priority list is generated; if the calculation is overtime, optimizing the structure of the decision model to obtain a simplified priority list; fusing the change trend of the physiological state of the patient with the dynamic adjustment model to obtain an updated action sorting sequence; and finally generating and outputting a personalized rehabilitation training scheme and outputting the personalized rehabilitation training scheme to a display device. According to the method, through multi-modal data fusion, tree-shaped decision risk assessment and real-time dynamic adjustment, the problems of insufficient personalized precision of a rehabilitation scheme and system response lag are solved, and the safety and suitability of training are improved.

The invention relates to a medical equipment fault prediction method and system based on multi-index fusion, and relates to the technical field of medical equipment fault prediction, and the method comprises the steps: collecting and associating multi-source heterogeneous operation data of target medical equipment; reversely analyzing and labeling event influence intervals in the time sequence dynamic monitoring data, and constructing a labeled time sequence data set of equipment state evolution; constructing a multi-task fusion prediction model; acquiring latest time sequence dynamic monitoring data in real time, inputting the latest time sequence dynamic monitoring data into the multi-task fusion prediction model, and synchronously acquiring an output future equipment comprehensive state index attenuation gradient prediction value and a discrete event occurrence probability prediction value; a fusion prediction score is calculated and a preventive maintenance alert is generated when a dynamically adjusted decision threshold is exceeded. The problems of high false alarm rate and insufficient early warning accuracy caused by the fact that traditional medical equipment fault early warning adopts a single judgment basis and related multi-class data are not integrated are solved.

The invention discloses a sewer pipe network defect detection method based on multi-label zero sample learning. The method comprises the following steps: generating defect description corresponding to each pipeline defect category through a large language model; performing feature extraction and domain adaptation on the pipeline inner wall image and the defect description by using a representation guide module to obtain global and local image features and defect detailed description text features; synthesizing global and local image features of the image, respectively calculating global and local matching scores of the global and local image features and detailed description text features, and fusing to obtain a defect initial prediction score; constructing a semantic relationship adjacency matrix for displaying the relationship between different defect categories; and correcting the initial prediction score by using the semantic relationship adjacency matrix between the categories to obtain a final prediction score of each defect category. According to the method provided by the invention, knowledge migration from known defects to unknown defects is realized by constructing a guidance-fusion-correction network, and the difficulty in identifying types of unseen defects is effectively solved.

The invention discloses a geographic data collection method and system based on machine learning, and relates to the field of geographic data collection, and the method comprises the steps: obtaining a historical collection record set of a target geographic region, carrying out the statistics of the historical collection record set, and obtaining the collection frequency distribution of different sub-regions in the target geographic region; according to the acquisition frequency distribution, identifying a historical low-coverage sub-region of which the acquisition frequency is lower than a first threshold value; generating a deviation correction guiding factor for each candidate acquisition position based on the spatial distribution characteristics of the historical low-coverage sub-region and the associated geographic element change characteristics; and in response to a current acquisition task request, gaining a prediction score according to the deviation correction guiding factor and information of a candidate acquisition position obtained from a pre-trained geographic element change prediction model. According to the method, a deviation correction mechanism is introduced, information acquisition efficiency and space coverage balance are cooperatively optimized in path planning, historical acquisition deviation is improved, and comprehensiveness and long-term application value of a geographic database are improved.

The invention discloses a genome variation cold and hot spot region prediction method and device, and the method comprises the steps: S1, carrying out the slicing of a target genome region according to a preset sliding window length, and constructing a multi-modal input tensor corresponding to a window; s2, inputting the multi-modal input tensor into a pre-trained deep learning prediction model, and outputting a cold and hot spot prediction score of each site in the window through a full connection layer; and S3, according to the cold and hot spot prediction scores, identifying a variation cold spot region and a variation hot spot region in the target genome region. According to the technical scheme provided by the invention, the dependence on the existing variation data density is eliminated, and non-blind area coverage in the whole exon group range is realized; meanwhile, the structured output based on the preset transcript coordinates can directly support clinical variation interpretation, a quantitative basis is provided for PM1 and cold spot evidence, and the proportion of unclear significance variation is effectively reduced.

The invention relates to a data quality checking method and device based on graph embedding, and belongs to the technical field of computers.The method comprises the steps that service data nodes and service domain nodes are input into a graph embedding module to obtain service data node vectors and service domain node vectors respectively; inputting the business data node vector and the business domain node vector into a feature conversion module, performing feature extraction, and then outputting an updated business data node vector and an updated business domain node vector; and constructing an edge vector of the updated service data node vector and the service domain node vector through a node aggregation module, performing feature extraction and prediction based on the edge vector, outputting a prediction score of an edge existing in the service data node and the service domain node, and determining a check result of the service data according to the prediction score. According to the method and the device, the automatic quality check of the irregularly input data is realized, the labor cost is saved, and the accuracy and the credibility of the data quality check are improved.

The present application relates to a bridge technical condition prediction method based on the fusion of graph neural network and time series modeling, belongs to the field of traffic infrastructure maintenance and artificial intelligence application technology, and solves the problems of insufficient utilization of space-time characteristics, lack of engineering logic constraints and rough prediction granularity of existing methods. The present application first collects historical technical condition data and pre-processes to obtain structured space-time sequence data; a double-branch feature extraction network extracts spatial correlation feature vectors and time series evolution feature vectors in parallel; a degradation trend prior feature vector is fused through a gating mechanism to generate a comprehensive feature vector; a double-branch continuous ordinal prediction head is used to output the initial continuous prediction score of the target bridge in the prediction year; finally, a time series consistency post-processing algorithm is used for logical constraint correction to generate the bridge technical condition grade prediction result. The present application effectively utilizes space-time characteristics and has strong engineering logic interpretability, and can realize continuous and accurate bridge technical condition prediction.

This invention discloses a multimodal collaborative denoising product recommendation method based on modality balance. The method first constructs behavior-aligned multimodal semantic encoding and projects it onto the behavior semantic space to obtain a multimodal feature product sequence. Next, for the multimodal feature product sequence, a multimodal-aware collaborative denoising module is constructed to collaboratively filter noise from each modality, resulting in a denoised intermediate product sequence. By introducing positional encoding in conjunction with the collaborative denoising module, an enhanced product sequence is obtained. Finally, based on the enhanced product sequence, cross-modal fusion weights are generated, prediction scores are calculated, and the product with the highest score is recommended. A joint loss function is constructed, and the global parameters are iteratively updated using a backpropagation algorithm. This invention effectively solves the problems of noise interference and modality learning imbalance in multimodal recommendation, suppresses the excessive dominance of strong modalities in the early stages of training, and improves the accuracy and robustness of product recommendations.

A method is disclosed. The method includes receiving a training dataset including a set of training samples. The method then includes obtaining a first parameter value and a second parameter value. After the parameters are determined, the training dataset can be fed into a machine learning model to train the machine learning model using a neighborhood-specific loss function. The method can then include receiving a second dataset including a set of second samples. Each second sample can then be input into the trained machine learning model to determine a prediction score for each second sample, and the prediction scores can form a bimodal distribution centered around the first parameter and the second parameter.

The invention discloses a multi-behavior recommendation method and device based on structure perception multi-view cascade fusion. The method comprises the following steps: loading a data set and constructing a global interaction graph; global embedding initialization is carried out, and global graph representation learning is carried out; behavior view decomposition is carried out for different behaviors; performing hierarchical fusion on different views and cascading historical behavior transmission information; the robustness is improved by using an auxiliary task of structure perception alignment; multi-task collaborative optimization training is carried out, and a main task based on personalized weighted Bayesian loss and an auxiliary task of structure perception alignment are optimized at the same time; and calculating a prediction score by utilizing the final embedding obtained by training and generating a recommendation list. According to the method, the problems of semantic collapse and static deviation of multi-view data can be effectively solved, and the accuracy and robustness of a recommendation system are improved.

The application discloses a news detection method and system based on a double-model collaborative optimization framework, and a medium, which comprises the following steps: screening news materials and news descriptions from a news material library to construct a training data set; inputting the training data set into a student model and a teacher network model respectively to obtain first image features and second image features, aligning the first image features to the space of the second image features to obtain third image features, inputting the third image features into a decoder of the teacher network model to perform target query optimization to obtain fourth image features, and taking the prediction score obtained by the teacher network model as a soft label to supervise the training of the student model until a preset condition is met to determine a trained target detection model; and inputting real-time acquired news materials into the target detection model to determine a news category and sensitive elements, so that the application can improve the accuracy and efficiency of news detection.

The invention discloses a climate prediction method and device based on hot start and background constraint and a medium. The method comprises the following steps: training an auto-encoder by using actual climate data in a first historical annual interval, and taking the auto-encoder as an initial climate prediction model decoder; obtaining first climate data of a current year and a second historical year interval predicted by a specified climate prediction model; for each month in the second historical annual interval, generating a sample pair of the month by taking the first climate data of the month and a preset number of months as an input feature and taking the actual climate data of the month as a label; the training sample library is used for training a climate prediction model trained in the last year with the minimum joint loss as the target, and the joint loss comprises a precision constraint representing model prediction performance, a climate background constraint and a prediction score constraint; the trained model is used for predicting the climate of the current year. The continuous evolution and stable prediction of the prediction model are realized, and the convergence speed of the prediction model is improved.

The invention relates to an enterprise tax risk prediction method and device, equipment and a medium. The method comprises the steps that firstly, a multi-tax-category original declaration data set is acquired, and a tax category network structure containing tax category nodes and display and implicit associated edges is constructed through a graph neural network; extracting node features, inputting the node features into a pre-trained random forest classification model, judging a matching state, and if deviation exists, calculating an initial risk probability score of each path; screening a high-risk path, generating a contradictory feature vector, calculating a quantitative risk assessment value, and fusing multi-dimensional indexes to obtain a tax risk prediction score; and finally, according to the score and a historical case feedback coefficient, adjusting an edge weight, constructing a map marking the weight and the conduction direction, and performing high-risk node screening, connection strength adjustment and association edge grading processing to obtain an optimized association map. The method breaks through the limitation of evaluation of a single tax category, effectively captures composite risks, and provides reliable support for refined collection and management and compliance management of enterprises.

The invention provides an application function recommendation method and device, and the method comprises the steps: constructing a text feature vector and a behavior feature vector based on application program data, and carrying out the fusion to obtain a fusion feature vector; determining an embedded vector corresponding to the fusion feature vector by using a heterogeneous graph information network model; inputting the embedded vector into a prediction model, and outputting prediction scores of the multiple functions; according to the prediction scores, the multiple functions are arranged in a descending order, and a preset number of functions with the highest ranking are selected as recommendation results to be output. By constructing a heterogeneous graph information network model with multiple types of nodes and multiple semantic edges and combining multivariate path mining, the accuracy, interpretability and generalization ability of processing a complex semantic relationship between a user and a function are improved. By capturing the attention weights of different relationships and enhancing the distinction degree of feature expression, the method can accurately model user requirements under the condition of data sparsity, and improves the robustness and performance of a recommendation system, thereby meeting the multi-dimensional requirements of different users.

The invention provides an electric vehicle charging pile recommendation method based on a heterogeneous information network. The method comprises the steps that historical charging POI page view data, charging pile scoring data of a user and charging plan search trend data and event search trend data from a search engine are collected and subjected to data preprocessing; constructing a heterogeneous information network; on the basis of the heterogeneous information network and the preprocessed data, the charging page view of the charging pile is predicted through a charging prediction model, and a charging page view prediction result is generated; based on the charging page view prediction result, through a collaborative filtering algorithm based on matrix decomposition, generating prediction scores of all unused charging piles by a user; the prediction score is corrected according to the weighted score of the POI around the charging pile, and a final recommendation score is obtained; and generating a charging pile recommendation list according to the final recommendation score. The accuracy and practicability of charging facility recommendation are improved.

An information processing system, which predicts an unknown binary relation between a treatment method and a biomarker based on a known ternary relation among the treatment method, the biomarker, and a disease, generates for each disease, based on the known ternary relation, a disease-specific bipartite graph that represents the binary relation between the treatment method and the biomarker, calculates, based on the disease-specific bipartite graph, a disease-specific inter-treatment-method similarity between treatment methods, a cross-disease inter-treatment-method similarity between the treatment methods, a disease-specific inter-biomarker similarity between biomarkers, and a cross-disease inter-biomarker similarity between the biomarkers, and calculates and outputs a disease-specific prediction score and a cross-disease prediction score of an unknown edge.

The invention provides an object sorting method and device, a model training method and device, a medium and equipment, and relates to the technical field of artificial intelligence, in particular to the technical field of machine learning and information retrieval. According to the implementation scheme, the method comprises the steps of obtaining a candidate list, a current search word corresponding to the candidate list and a historical behavior sequence; based on the historical behavior sequence and the current search word, utilizing a first interaction network in a sorting model to obtain a first fusion feature; for each candidate object in the candidate list, based on the object information of the candidate object and the first fusion feature, predicting to obtain a prediction score of the candidate object; and sorting the candidate objects in the candidate list based on the predicted score of each candidate object in the candidate list.

This invention discloses a recommendation method and system that integrates content similarity and a neural collaborative filtering model, belonging to the field of artificial intelligence technology. The technical problem it addresses is how to overcome the low recommendation accuracy and poor adaptability in scenarios with sparse data and cold starts. The method includes: merging project attribute data and user-project interaction data corresponding to the same project; constructing a project feature matrix based on extracted project feature vectors; calculating the mean vector of all project feature vectors as a user profile vector; constructing a sample set based on positive and negative samples; building a recommendation model based on a neural collaborative filtering model, and training and testing the recommendation model based on the sample set; calculating a matching score for each project based on the similarity between the user profile vector and the project feature vector, and predicting the interaction probability between the user and each project using the trained recommendation model as a prediction score; and weightedly fusing the matching score and the prediction score.

The application provides an information recommendation method, system, device and storage medium, comprising: by associating a special graph structure for different expert networks, the physical and logical decoupling of modal information is realized, and the embedding feature information of the target user and the candidate item in different semantic dimensions can be more accurately described. By calculating the data routing distribution with the user-item interaction instance as the granularity, and combining the progressive routing strategy with the prior routing distribution to obtain a hybrid routing distribution, accurate expert scheduling can be realized in different context environments, that is, it can automatically identify which modal features play a key role in user decision-making in a specific scenario, thereby accurately capturing the fine-grained preferences of users. Through the gating selection mechanism, the feature contribution of key experts is retained, effectively filtering the noise interference generated by irrelevant modalities, and further improving the accuracy of the prediction score and the reliability of the recommendation result.

This invention discloses a gradient-guided feature activation-based self-enhanced method for single-class target perception, comprising: determining the target category to be detected in a target image; inputting the target image into a target detection network for a first forward propagation to obtain the prediction score of the target category corresponding to the prediction layer of the target detection network; performing gradient backpropagation on the prediction scores to obtain the gradient information of the feature map of the target image corresponding to each prediction layer; determining the channel weights and correlation coefficients of the feature map corresponding to each prediction layer based on the gradient information; determining the final channel weights of the feature map based on the channel weights and correlation coefficients; weighting the feature map corresponding to the prediction layer using the final channel weights to obtain the enhanced feature map; and inputting the enhanced feature map into the target detection network for a second forward propagation to obtain the enhanced target detection result. Thus, without retraining the network, the method improves the ability to recognize specific target categories while reducing the additional cost of retraining the network.

The invention relates to a method for realizing GPCR receptor target drug discovery based on a model consensus mechanism. The method comprises the following steps: constructing composite training data of a general data set and a special data set; establishing a cascade subtask model, and sequentially performing interaction prediction, combination strength prediction and functional activity prediction; capturing interaction characteristics through a double-view-angle coding and interaction module, and outputting a standardized prediction score; generating unified confidence and consistency measurement for interaction prediction, affinity prediction and functional activity prediction tasks, and obtaining a final screening result. According to the method, the device, the processor and the computer readable storage medium for realizing GPCR receptor target drug discovery based on the model consensus mechanism, through cascade multi-task modeling and multi-scale feature fusion, layered inference of protein-small molecule action is realized, single-model deviation is reduced, prediction precision is improved, and the method and the device are suitable for large-scale popularization and application. Unstable prediction is automatically screened out by using a model consensus mechanism, and the robustness and generalization ability of a result are improved.

The application discloses a kind of monitoring analysis evaluation methods applied to urban water supply, it is related to water supply evaluation technical field, and the present application establishes environmental factor library and characteristic index library based on historical evaluation record in water supply supervision platform;The operation data of characteristic index in historical evaluation record is collected, and is combined with the numerical value of artificial evaluation result of historical evaluation record, to generate score prediction model;The data of target environmental parameter is collected, and is combined with score difference, to generate score threshold factor model;The abnormal time interval of two adjacent abnormal records is calculated, and is combined with score threshold factor and prediction score, to generate prediction time interval model;Real-time acquisition data in environmental factor library and characteristic index library, calculate real-time prediction score and real-time score threshold factor, generate corresponding evaluation strategy, improve water supply system safety and management efficiency.

The application relates to the field of artificial intelligence and computer vision, and provides a three-dimensional point cloud scene target positioning method based on a natural language instruction, which comprises the following steps: screening a relation triple containing an object category in a three-dimensional visual scene; determining a hierarchical attribute of the object category and constructing a hierarchical knowledge graph; analyzing a given natural language description to obtain a plurality of two-order semantic triples; adding the hierarchical attribute to the object in the two-order semantic triple to obtain a hierarchical two-order semantic triple; determining an initial visual feature of an object in a three-dimensional point cloud scene to be recognized; outputting a prediction score of the object; screening the objects corresponding to the prediction scores from high to low to obtain candidate objects; outputting a matching score of the candidate objects and a first high-order semantic triple; and determining a target object in the three-dimensional point cloud scene to be recognized. The application improves the prediction accuracy and has stronger robustness.

The application discloses a large language model test learning method, device, equipment and medium, test data is input into a large language model with low rank adaptive parameters, the prediction score of each token in the vocabulary at each generation step is obtained; the token-level evidence quality of the generation step is extracted based on the prediction score, and the token-level cognitive uncertainty of the generation step is calculated; the token-level cognitive uncertainty is smoothed, the stable area of the generation step is determined according to the smoothed uncertainty, and the mask of the model update window is constructed; the mask entropy loss is calculated based on the mask, and the low rank adaptive parameters are updated through back propagation. The application can solve the problem of adaptive drift in model updating in related technologies.

The invention belongs to the technical field of artificial intelligence and meteorological prediction, and provides a meteorological mid-term forecast enhancement method and system based on a space-time condition generative model, and the method comprises the steps: carrying out the preprocessing of numerical weather forecast grid data and meteorological station observation historical data, and obtaining a forecast feature sequence aligned with the position of a meteorological station; on the basis of the forecast feature sequence, combining time information with site geographic information to generate a spatio-temporal context vector representing future weather evolution; generating a random potential variable carrying reasonable uncertainty based on the context vector; and taking the context vector and the random potential variable as input, predicting the deviation of numerical weather forecast in a residual learning mode, adding the deviation and an original numerical forecast result, and outputting an optimized and corrected final forecast value of the station meteorological elements. According to the method, the reconstruction error, the potential spatial distribution and the probability forecast score are jointly optimized, and the problems of mode collapse and fuzziness easily occurring when diversity data is generated by a traditional model are solved.

This application discloses a method, device, and storage medium for serial number recognition, relating to the field of image recognition technology. The method acquires a serial number image, inputs it into a convolutional neural network (CNN), and uses the CNN to increase the dimensionality of the image's channels to obtain a feature map. The feature map is then input into a bidirectional long short-term memory (LSTM) network to determine the corresponding bidirectional temporal feature sequence. This bidirectional temporal feature sequence is then input into a linear mapping layer to determine the prediction score for each character category. Based on the prediction score, the corresponding serial number character is determined. This method can adapt to variations in the format and length of different serial number types, thereby improving the generalization ability of serial number recognition.