154 results about "Domain identification" patented technology

The invention relates to the technical field of remote sensing image processing, and discloses a remote sensing image adaptive identification method and system for territorial space planning, and the method comprises the steps: obtaining a multi-source remote sensing image data set of a research region, feature extraction, cloud detection, quality evaluation and adaptive preprocessing are carried out; carrying out prototype network coding, calculating a category prototype and probability, and supporting fine tuning of a set; carrying out multi-scale cavity convolution and category scale attention fusion; evaluating the adaptability score of the comprehensive fusion feature map set, and carrying out weighted fusion, classification and normalization; change detection is carried out, stable and change regions are segmented, and time sequence context features are extracted and constrained optimization is carried out; entropy is fused, a boundary is decided, uncertainty is estimated, and weighted fusion is carried out according to a change area; conditional random field optimization, confidence level grading and connected domain identification are carried out; the automation level, the adaptive capacity and the recognition reliability of remote sensing monitoring of territorial space planning are improved.

The invention relates to an abnormal transaction behavior recognition method and system based on artificial intelligence, and belongs to the technical field of financial transaction risk control, and the recognition method comprises the steps: collecting real-time transaction flow data, user behavior time sequence data and association graph data of a user, carrying out the space-time alignment processing, and generating a space-time aligned structured data set; dynamically calculating a transaction time attenuation factor; extracting a composite feature vector, inputting the composite feature vector into a pre-trained dual-channel decision model, and performing confidence weighted fusion on a dual-channel decision result based on a time decay factor to obtain a final risk score; dynamically adjusting a risk judgment threshold, and generating a dynamic decision boundary; judging whether the final risk score exceeds a dynamic decision boundary, if so, outputting an abnormal transaction behavior recognition result, and judging a transaction risk level; and triggering a risk disposal action corresponding to the transaction risk level. The method can improve the understanding depth of complex transaction behaviors, and reduces the false alarm rate of abnormal transaction recognition.

The invention relates to the technical field of big data analysis, and discloses a prospecting target area positioning method and system based on big data, and the method comprises the steps: collecting multi-source exploration data in real time through distributed nodes, completing coordinate normalization, semantic alignment and time synchronization through a spatial heterogeneous data flow engine, and generating a standardized incremental data block; performing local feature sensitivity analysis based on the historical model library, identifying a newly added feature dimension, and performing parameter increment updating by adopting a sliding window gradient descent method; inputting the updated model into a target evolution model driven by a Bayesian space-time probability field, and dynamically calculating the metallogenic probability of each space grid in combination with a stress field, an element migration path and historical verification data; and generating high, medium and low three-level target area maps according to probability sorting, and pushing the high, medium and low three-level target area maps to a three-dimensional visual decision terminal. According to the method, minute-level dynamic response of the target region under triggering of newly-added data is realized, computing resource consumption is reduced to be less than 5% of that of an original system, and prospecting efficiency and abnormal region identification timeliness are improved.

Systems and methods of processing domain-specific content in a generative AI system including receiving a prompt, tokenizing the prompt, identifying an identified domain of the tokenized prompt identifying domain-specific functions within the identified domain, generating domain-specific sub-functions from the domain-specific functions according to a hierarchical mapping, generating H-Tokens, each encapsulating one of a domain-specific function or a domain-specific sub-function and relationships domain-specific functions and the domain-specific sub-functions, implementing the of H-Tokens, assembling a response using the implemented of H-Tokens, and transmitting the response to the user.

The invention discloses a spatial domain identification method and device based on multi-modal topological consistency, and belongs to the field of transcriptome spatial domain identification, and the method comprises the steps: constructing a multi-layer network which is in one-to-one correspondence with modal information contained in a biological tissue based on spatial transcriptomics data of the biological tissue; extracting a consensus structure feature shared by the multi-layer network and a specific structure feature specific to each layer of network; constructing a cell consistency network of the biological tissue according to the consensus structural features, the specific structural features and the multilayer network; and performing clustering processing on the cell consistency network to obtain recognition results of different spatial domains in the biological tissue. According to the method, the heterogeneity problem among different modal data can be overcome, and the spatial domain recognition effect is better.

The invention relates to a customer service scene-oriented collaboration system and method, and belongs to the technical field of artificial intelligence. In the system, a shared memory storage center stores knowledge entries from a plurality of customer service agents in a predefined data structure; the shared memory storage center is provided with an access control mechanism associated with a visible domain identifier, and an access range is limited during retrieval through a logic partition; the memory management module receives candidate knowledge entries from each customer service agent based on a trigger event, determines visible domain identifiers for the candidate knowledge entries, associates the candidate knowledge entries with the visible domain identifiers, and stores the associated candidate knowledge entries and the visible domain identifiers in the shared memory storage center; the problem processing and decision-making module responds to an access request from any customer service agent and determines a corresponding visible domain identifier according to the identifier of the customer service agent; and executing access based on the visible domain identification and the input question, and retrieving knowledge entries with access from a shared memory storage center to realize cross-agent multiplexing.

The invention discloses a violation person identity recognition method for gas turbine power plant on-site safety supervision, relates to the technical field of person identity recognition, and aims at recognizing specific violation behaviors, positioning the occurrence time and space of the violation behaviors and marking corresponding behavior subjects. The method comprises the following steps: acquiring operation image data, extracting behavior subject image features and equipment acquisition data, generating fusion identity information, matching violation identities, outputting structured records, evaluating violation risk levels, verifying the identities of persons in charge and filing violation records, and automatically identifying dangerous behaviors in a gas turbine power plant through the operation image data acquisition and violation behavior identification model. And in combination with the dynamic attitude vector sequence and equipment acquisition data, generating identity feature fusion data, performing map matching, identifying the identity of a violation person, labeling through a risk level label, pushing to a responsible person terminal, completing identity confirmation and violation behavior recording and archiving, and realizing whole-process supervision and closed-loop management.

The invention discloses a product production process scheduling method and system based on intelligent manufacturing, and belongs to the technical field of intelligent manufacturing. According to the method, rapid local rescheduling is realized through fault influence domain identification, a three-level response strategy and real-time equipment health assessment. When an equipment fault signal is detected, a to-be-executed task on the fault equipment is identified, a subsequent task chain is traced through the technological process dependency relationship, the boundary of a fault influence domain is determined, and the influence domain relates to 10%-20% of tasks in the whole production plan. And selecting a corresponding response strategy according to the predicted repair time, adopting a task delay strategy when the predicted repair time is less than 30 minutes, performing local redistribution on tasks in the influence domain when the predicted repair time is less than 30 minutes to 2 hours, and expanding the influence domain to the whole process for comprehensive optimization when the predicted repair time is more than 2 hours. Meanwhile, monitoring data such as equipment vibration, temperature, current and rotating speed are collected to calculate a health degree score, and 5%-10% of productivity redundancy is reserved for high-risk equipment.

The invention discloses a deep learning-based rock hardness intelligent identification model and method, a DF (data fusion)-improved CNN (Convolutional Neural Network) model provided by the invention still keeps 97.92% stable accuracy under continuous cutting and different working conditions, and the problems of tedious artificial feature extraction engineering and weak model generalization in traditional rock hardness identification are solved. And the problems of insufficient characterization capability and weak model generalization performance of a traditional method are solved. Different from a feature screening process dominated by expert experience in a traditional mode, the method provided by the invention realizes feature adaptive extraction through a multi-channel convolutional neural network, carries out overlapped sampling data enhancement on an original vibration signal, constructs a time-frequency entropy multi-domain fusion graph through short-time Fourier transform, and obtains a time-frequency entropy multi-domain fusion graph; experimental verification shows that the classification accuracy of the multi-channel structure is greatly improved compared with that of a single-channel model, and time-frequency entropy multi-domain fusion is more accurate than that of a single-domain recognition model.

The invention discloses a robust radio frequency fingerprint identification method based on Barlow Twins domain adaptation. Aiming at the problems of training and deployment environment distribution offset and cross-domain identification performance reduction caused by wireless channel multipath and time-varying characteristics in the prior art, the invention provides a domain adaptation framework combined with feature decoupling. The method comprises the following steps: firstly, converting a received time domain signal into a short-time Fourier transform spectrogram to reserve a time-frequency structure; then, a double-end convolutional encoder is constructed, bottom layer features are extracted through a shared backbone network, and device fingerprints and channel features are obtained in an identity branch and a channel branch respectively; in order to realize thorough decoupling of the two types of features, Barlow Twins-based independence regularization is introduced, and the identity features and the channel features are statistically orthogonal by minimizing a cross-correlation matrix of the identity features and the channel features, so that purer fingerprint features are obtained, channel interference is effectively eliminated, and generalization and recognition precision of the model under an unknown channel are remarkably improved.

The invention discloses a spatial transcriptome data spatial domain identification method based on multi-space self-supervised contrast learning, and the method comprises the steps: carrying out the modeling to generate a spatial neighborhood graph, keeping the graph structure unchanged, disorganizing node features, and carrying out the data enhancement, thereby obtaining an enhanced graph; constructing an encoder based on a graph neural network, extracting spatial transcriptome data fused with spatial information and gene information to obtain potential embedding, and sending the potential embedding into a multi-space generator to generate multiple groups of rich graph feature representations; fusing graph feature representation and potential embedding to obtain refined representation, reconstructing a gene expression matrix through a decoder, and adding contrast learning loss and reconstruction loss as a total objective function; and updating network parameters by adopting an Adam optimizer according to the obtained total objective function to complete spatial transcriptome spatial domain identification. Spatial transcriptome data are fully mined from global and local angles, and accurate spatial domain identification is realized.

The invention relates to the technical field of image recognition and foam detection, and discloses a surfactant foam amount prediction method and system based on vision, and the method comprises the steps: collecting a foam image, carrying out the region extraction through the combination of guided filtering and a self-adaptive threshold value, extracting a small foam structure feature vector through employing Euclidean distance transformation and a watershed algorithm, and carrying out the prediction of the foam amount of a surfactant. And constructing a foam quantity prediction model, and outputting a foam quantity result. In the prior art, a method based on area statistics or connected domain identification is low in identification precision under a high-density small bubble scene, and especially under the conditions that the boundary of a foam region is fuzzy and small bubbles are seriously adhered, accurate bubble diameter extraction is difficult to realize. Due to the fact that guided filtering enhancement, a watershed segmentation method and image and structure double-branch prediction are introduced, accurate segmentation of a dense small bubble area and prediction of the foam amount are achieved, and the accuracy and adaptability of foam amount prediction in a non-ionic surfactant complex foam scene are improved.

The invention provides a visual question and answer method based on a domain self-adaptive retrieval decision, which comprises the following steps of: forming an input triple (x, q, d) comprising an image, a question text and an image description text; feature modal extraction and domain identification; generating an explicit reasoning track and a preliminary answer by using a chain reasoning technology CoT; according to a preset decision rule, judging that the preliminary answer is output as a final answer or enters the next step; based on the input triad (x, q, d) and the reasoning track, image retrieval and text retrieval are executed, and an enhanced knowledge set is generated; using the enhanced knowledge set to generate a final answer through a chain reasoning technology CoT, performing credibility verification, and outputting the final answer or a preset unknown identifier according to a verification result; according to the method, through reasoning-driven adaptive retrieval and multi-modal knowledge reordering, efficient utilization and real-time supplement of external knowledge are realized, and the accuracy and robustness of visual questions and answers are effectively improved.

The invention relates to a transferable vector quantization alignment method and device based on unsupervised domain adaptation, and the method comprises the steps: extracting the features of a source domain and a target domain of a source domain data set and a target domain data set through a feature extraction unit, and calculating an overall feature distribution loss function through searching a feature item closest to the features in a codebook; calculating a local alignment loss function through a bottleneck layer, performing classification through a classifier to obtain a source domain pseudo-label and a target domain pseudo-label, calculating a cross entropy classification loss function according to the source domain pseudo-label and a corresponding truth value label, performing normalization processing on the target domain pseudo-label, introducing mutual information to obtain a sample weight in each target domain data set, and obtaining a sample weight in each target domain data set; and calculating mutual information weighted maximization confusion matrix loss functions, and updating parameters in each module by using the loss functions until convergence to obtain a target classification architecture with cross-domain extraction, feature alignment and time sequence signal classification capabilities. By adopting the method, the identification performance of the label-free target domain can be improved.

The invention relates to the technical field of computer vision and artificial intelligence, in particular to a cross-domain three-dimensional perception method combining geometric and semantic dual paths, which comprises the following steps of: 1, acquiring a multi-view image, and extracting a multi-scale two-dimensional feature map by using a two-dimensional feature encoder; 2, inputting the extracted two-dimensional feature map into a geometric generalization path, and carrying out the operation of self-supervised depth enhancement and explicit fusion to obtain geometric enhancement features; 3, inputting the extracted two-dimensional feature map into a semantic generalization path, and carrying out semantic query and domain confrontation fusion operation to obtain semantic enhancement features; and 4, finally fusing and decoding the geometric enhanced features and the semantic enhanced features to obtain a three-dimensional occupancy graph. The generalization performance of the method is improved in a breakthrough mode, and the method has the advantages that the practicability of the model is greatly enhanced, geometric perception is more accurate, three-dimensional reconstruction is more fidelity, semantic understanding is more consistent, and cross-domain recognition is more reliable.

The invention provides an Android malicious software detection method based on a multi-task learning mechanism, and belongs to the technical field of mobile security and malicious software detection. Comprising the steps of 1, extracting malicious family correlation, and extracting correlation among different malicious families through a multi-task learning mechanism by adopting a deep learning model based on a CNN (convolutional neural network) and a Bi-LSTM (bidirectional long short-term memory) network; 2, generating a malicious space, and depicting a high-risk sample region through feature space generation, threat region identification and region priority ranking; 3, identifying high-risk samples, and screening the samples based on region priorities and position attributes; and step 4, decision optimization: carrying out fine adjustment on the detection model by combining high-risk samples and known sample anchor points so as to improve the detection capability of unknown malicious families. According to the method, through multi-task learning and malicious space analysis, the detection effect of the method on unknown Android malicious software is improved.

The invention relates to the technical field of region segmentation, in particular to a splicing type prefabricated part splicing region identification method based on image semantic segmentation, which comprises the following steps of: acquiring a part image and extracting a light and shade change region, analyzing pixel arrangement and a jump edge, endowing a semantic label, tracking a contour path and extracting a trend and a direction. Recognizing turn-back, eliminating abnormal edge segments, recombining paths, extracting center line coordinates, adjusting directions, complementing point positions, and generating a label map to divide a splicing seam area. According to the method, boundary regions are divided through pixel arrangement and brightness edge connectivity, independent labeling of semantic blocks is completed, a structured edge segment set is constructed through path trend and turning point information, the coherence and connection relation sequence of the boundary direction is kept, direction sudden change positions are recognized, path anomalies are eliminated, and structural fracture and overlapping are reduced; central line pixels are extracted, a path trunk is formed through point supplementing operation at the turning position, an attribution area is delimited by combining contour extension and label coding, and the consistency and segmentation integrity of the abutted seam boundary are improved.

The application discloses a mixed layer thickness identification method based on a carrier coordinate system, and the identification method comprises the following steps: importing flow field data; uniformly processing the flow field data, and marking fluid domains and non-fluid domains; identifying active flow and passive flow, and respectively tracking characteristic flow lines of the active flow and the passive flow to determine the boundary of a mixed area; establishing a carrier coordinate system, and determining an integral path set of the mixed area; integrating all the integral paths to obtain the mixed layer thickness of each integral path, i.e., to obtain the mixed layer thickness along the flow direction of the mixed area; screening out integral path discrete points with abnormal growth or reduction of the mixed layer thickness before and after the mixed layer thickness, and removing the integral path discrete points; drawing a curve graph of the mixed layer thickness along the flow direction, and completing the identification of the mixed layer thickness. The application is applied to the field of flow field mixing, can meet the calculation requirements of the mixed layer thickness under more general flow field conditions, and provides more accurate data for flow field analysis, mixed supercharging evaluation and other applications.

A cross-domain recognition and drift compensation method for electronic noses based on contrastive variational autoencoders (CVAEs) is proposed, belonging to the interdisciplinary field of electronic nose systems and artificial intelligence. The method involves constructing a labeled dataset, building an initial C-VAE network and establishing a basic loss function, connecting the initial C-VAE network with a classifier to form a joint learning network, introducing a classification loss function to construct a combined C-VAE loss function, and training to obtain an initial model. A target domain dataset is then constructed and input into the initial model to obtain the latent space vector of the target domain. Mean-variance transformation is used to achieve style transfer and data augmentation of source domain features. The latent feature structure of the target domain is optimized through various constraints. A joint training set is constructed, and the depth alignment of features between the source and target domains is achieved by constraining the distribution differences of features between the two domains. Training is performed under the joint loss function, and progressive scheduling is applied to obtain a cross-domain gas recognition model. This method effectively solves the nonlinear drift problem of electronic nose sensors.

The invention discloses a low-voltage topology identification method, system and device based on characteristic current real-time sampling analysis and a medium, and relates to the technical field of data processing, the identification method comprises the following steps: collecting sampled current signals in real time through a master control intelligent terminal, and carrying out grouping detection; dynamically selecting three detection points on a current signal propagation path, calculating a distribution characteristic value based on a spatial position coordinate, and generating a topological matching correction coefficient K; according to the correction coefficient K, dynamically adjusting a fault-tolerant threshold value matched with the sequence, and if an adjusted matching result meets a condition, generating an identification record representing that a connection relationship exists between the current sending node and the sampling node; and recursively sorting the hierarchical relationship of all the generated identification records according to the number of identification times of the nodes from low to high until identification and redisk of all the nodes are completed, and generating and storing a final topological structure. According to the invention, the operation and maintenance efficiency and the intelligent level of the power grid are improved.

The invention discloses a Sentinel-2 remote sensing image water body region identification method based on an attention mechanism, and the method comprises the following steps: (1) obtaining and preprocessing a Sentinel-2 multispectral image, cutting to generate a multi-scene sample containing a mountain area fine river, a city point-shaped water body, cloud interference, an island large-range water body and the like, and constructing a water body region data set; (2) embedding coordinate attention (CA) in the U-Net semantic segmentation network; (3) constructing an attention-enhanced water body region segmentation network; training is carried out on a data set, and the model is evaluated through MIOU, MPA, Recall and Precision; and (4) verifying the Sentinel-2 multi-complex scene image by adopting a U-Net + CA model to obtain a water body region segmentation result. According to the method, the recognition precision of the small water body area and the water body area under the complex background can be remarkably improved, the boundary is smoother, and the method has high robustness to cloud and mist and shadows and is suitable for flood disaster monitoring and water resource management.

The application discloses a training method of a property customer service question and answer model, and comprises the following steps: inputting a multi-stage knowledge mask strategy into a BERT model for pre-training to obtain different language granularity information; and performing domain recognition training based on the language granularity information to obtain the property customer service question and answer model. According to the application, the multi-stage knowledge mask strategy and the domain recognition are trained in the BERT model, so that the user's question can be accurately understood, and the user experience is improved.

The invention discloses a long document FAQ list generation method and system based on a large model and a storage medium. The method comprises the steps of performing text preprocessing and fragment division on an input long document, determining a field to which a text fragment belongs in combination with a field identification model, automatically generating candidate questions based on a trained language model, inputting the candidate questions and the corresponding text fragment into an extraction type question and answer model to extract answer fragments, and performing question and answer extraction. The trained language model is combined with the context to complement answer segments to generate answer texts, question and answer pairs are formed, correlation scores are calculated through semantic similarity, keyword coverage and confidence, the question and answer pairs are subjected to duplicate removal and sorting, and an FAQ list is generated. According to the method, high-quality question and answer extraction of cross-domain long documents is realized by introducing a large-scale pre-training language model, and the question and answer generation accuracy, context consistency and FAQ sorting efficiency are remarkably improved.

The application provides a BART-based multi-task semantic parsing model, and belongs to the technical field of natural language processing, the model comprises a word embedding layer, a BART coding layer, a domain classifier, a BART decoding layer, a probability decoder, a SPARQL decoder and a grammar checker, the seven parts are matched, and based on the Encoder-Decoder model architecture of the Transformer, text noise is increased through means such as word deletion, sentence permutation transformation, document rotation and word filling, and the noise-bearing input decoding is mapped to the original text, a sequence-to-sequence encoder is obtained through training, and better effects are achieved in generation tasks such as question answering, translation and abstracting. The application directly converts the natural language into the knowledge graph query language SPARQL, simplifies the question answering step to reduce error accumulation, and identifies the domain of the question, queries the corresponding professional domain knowledge base according to the domain, and thus the question answering accuracy is improved.

The invention particularly relates to a multi-slice spatial transcriptome spatial domain identification method, which comprises the following steps of: preprocessing spatial transcriptome data of a plurality of slices, constructing an unweighted undirected graph based on spatial coordinates, and modeling each slice as an attribute graph; constructing a combined learning model fusing a graph attention auto-encoder and an adversarial module, extracting node low-dimensional embedded representation from gene expression and a spatial neighborhood by using a graph attention mechanism, and guiding the encoder to extract cross-slice shared feature information through an adversarial training strategy based on a Wasserstein distance; in order to improve the adaptability of the model to large-scale graph data, a sub-graph training mechanism based on neighbor sampling is introduced; and clustering the embedded representation to realize the joint identification of the multi-slice spatial domain. According to the method, effective information in the slices and between the slices is utilized, specific structures of different slices are identified while a common spatial domain is accurately identified, the training efficiency is high, and the method is suitable for a conjoint analysis task of large-scale spatial transcriptome data.

Embodiments of the present application provide a model training method and device, electronic equipment and storage medium, and relate to the technical field of computers. The method comprises: retrieving a set of documents from a preset document library based on a set of keywords; clustering the set of documents to obtain a plurality of subsets of documents; extracting at least one document from each subset of documents as a first document; determining positive samples and negative samples from the first document through a plurality of large language models and training an initial model to obtain a target model. After training the baseline model on the few-sample labeled dataset of the multi-model voting, the baseline model is enhanced using an active learning strategy. First, the baseline model is used to evaluate a randomly selected set of documents, and the sample with the lowest prediction is iteratively annotated. Using the large model debate idea, when the accuracy, recall rate and other indicators of the model no longer have significant improvement, the iteration is stopped, so that the trained model has high target field recognition ability.

The application relates to a spatial transcriptome spatial domain identification method and system based on double contrast learning. The method comprises the following steps: performing Gaussian smoothing processing on gene expression data to obtain smoothed gene expression data; constructing an initial adjacency matrix based on spatial coordinate data, performing edge pruning to obtain an optimized adjacency matrix; inputting the smoothed gene expression data and the optimized adjacency matrix into a graph convolution network encoder to obtain low-dimensional latent features and reconstruction features, and calculating a reconstruction loss; inputting the low-dimensional latent features into a multi-branch multi-layer perception fusion module to obtain enhanced features; constructing a regional contrast learning task, calculating a regional contrast loss, constructing a topological contrast learning task, and calculating a topological contrast loss; optimizing model parameters through a total loss, using an optimized model to perform spatial domain identification, and obtaining a spatial domain identification result. The spatial domain identification accuracy can be improved and the cost can be reduced.

The application relates to a pedestrian attribute recognition method based on deep learning and belongs to the technical field of pedestrian recognition. The recognition method comprises the following steps: constructing a pedestrian attribute recognition network; the pedestrian attribute recognition network comprises a backbone network, a first pooling layer and a full connection layer, and an attention module is introduced between the backbone network and the first pooling layer; training the pedestrian attribute recognition network; and performing pedestrian attribute recognition according to the trained pedestrian attribute recognition network. The pedestrian attribute recognition network for performing pedestrian attribute recognition introduces an attention module between the backbone network and the first pooling layer, the importance of high-level semantic information is distinguished through the attention module, and then the high-level semantic information is purposefully selected and processed by the first pooling layer, the global features corresponding to the high-level semantic information are emphasized, and the recognition accuracy is improved. The pedestrian attribute recognition network structure is simple, the complexity of the network structure is avoided from being greatly increased, and the recognition efficiency is improved.

The invention discloses a judicial text entity relationship joint extraction method based on regional vertex labeling. The method comprises the following steps: coding judicial text cleaning data through a pre-training language model BERT to obtain corresponding vector representation and embedded representation; a head entity and a tail entity in a triple corresponding to the original judicial text data form a rectangular area in the target representation set, and four vertexes of the rectangular area are identified to identify the triple; calculating probability scores that target representations in the target representation set belong to the four vertexes; and extracting the triple of the original judicial text data by combining the loss function with the character pairs with the distributed labels. According to the method, the key information in the original judicial text data is converted into the formatted triple, the core relationship between the named entity and the positioning entity in the judicial text is accurately identified, the structured representation of the irregular text is realized, and judicial workers are further assisted in understanding the case.

Implement measures in areas where they would be most effective if implemented. [Solution] A prediction device accepts the designation of a specific region where a measure related to the provision of a service will be implemented, identifies a precedent region where the measure has been implemented in the past, and predicts whether potential member users in the specific region who have not yet used the service will use the service after the measure has been implemented. Regional similarity, which is the degree of similarity between the precedent region and the specific region, and user similarity between member users in the precedent region who had not used the service before the measure was implemented and member users in the specific region, are input into a machine learning model to predict whether or not the service will be used. The machine learning model is trained by using as training data data correlating the number and usage rate of the service in the precedent region with the regional similarity between the precedent region and past regions where measures were implemented before the precedent region, and the user similarity between member users in the past region and the precedent region.