Abstract: A training method performed through batch learning by a computer includes: obtaining training data including first time-series data and second time-series data different from the first time-series data; performing first training processing of training a neural process (NP) model, which outputs, using a stochastic process, a prediction result that takes uncertainty into account, to predict first and second time-series data distributions, based on the first time-series data and second time-series data; and performing, using a contrastive learning algorithm, second training processing of (i) training the NP model to bring close to each other first sampling data items generated by sampling from the first time-series data distribution, (ii) training the NP model to bring close to each other second sampling data items generated by sampling from the second time-series data distribution, and (iii) training the NP model to push away the first and second sampling data items far from each other.

Abstract: A demand prediction method using a computer includes calculating an error between a prediction value of an amount of a past demand and a measurement value of the amount of the past demand, the prediction value of the amount of the past demand being calculated by inputting a measurement value of a past explanatory variable to a prediction model that is constructed in accordance with a measurement value of an amount of a demand at a predetermined location and an explanatory variable that serves as an external factor that affects an increase or a decrease in the amount of the demand at the predetermined location, determining whether the calculated error is an abnormal value, acquiring a new explanatory variable if the error is determined to be the abnormal value, updating the prediction model in accordance with the acquired new explanatory variable, and newly predicting an amount of a future demand using the updated prediction model.

Abstract: A method is provided for reducing computer memory usage and processing time in assessing effectiveness of a prediction model. The method includes acquiring a plurality of records and sorting the records based on prediction values. The method also includes determining two of three performance data parameters, including a number of selected records and a number of correct decisions, and determining a level of effectiveness of the prediction model based on the two of three performance data parameters. The method further includes generating a graph illustrating the level of effectiveness of the prediction model based on the two of three performance parameters, and visually conveying the level of effectiveness of the prediction model utilizing the generated graph.

Abstract: An anomaly detection method includes: extracting, for each of a plurality of learning packets obtained, all possible combinations of N-grams in the payload included in the learning packet; counting a first number which is the number of occurrences of each combination in the payloads of the learning packets; calculating, as anomaly detection models, first probabilities by performing smoothing processing based on a plurality of the first numbers; and when the score calculated for each of a plurality of packets exceeds a predetermined threshold that is based on the anomaly detection models stored in a memory, outputting information indicating that the packet having the score has an anomaly.

Abstract: An anomaly detection method includes: extracting, for each of a plurality of learning packets obtained, all possible combinations of N-grams in the payload included in the learning packet; counting a first number which is the number of occurrences of each combination in the payloads of the learning packets; calculating, as anomaly detection models, first probabilities by performing smoothing processing based on a plurality of the first numbers; and when the score calculated for each of a plurality of packets exceeds a predetermined threshold that is based on the anomaly detection models stored in a memory, outputting information indicating that the packet having the score has an anomaly.

Abstract: A method includes estimating a model parameter in a case where operating states of plural electric devices are modeled by a probability model by using a total value of power consumption of the plural electric devices connected with a panel board. In the estimating, the model parameter in which likelihood calculated by a likelihood function becomes a maximum is estimated based on characteristics of power data that may be predetermined as prior knowledge from an operation tendency of each of the plural electric devices, the probability model is a factorial hidden Markov model (FHMM), and the likelihood is a value that indicates certainty of a pattern of a total value of the power consumption, which is modeled by the FHMM, of the plural electric devices with respect to a total value of the power consumption that is actually measured.

Abstract: A demand prediction method using a computer includes calculating an error between a prediction value of an amount of a past demand and a measurement value of the amount of the past demand, the prediction value of the amount of the past demand being calculated by inputting a measurement value of a past explanatory variable to a prediction model that is constructed in accordance with a measurement value of an amount of a demand at a predetermined location and an explanatory variable that serves as an external factor that affects an increase or a decrease in the amount of the demand at the predetermined location, determining whether the calculated error is an abnormal value, acquiring a new explanatory variable if the error is determined to be the abnormal value, updating the prediction model in accordance with the acquired new explanatory variable, and newly predicting an amount of a future demand using the updated prediction model.

Abstract: A data analysis method decomposes a fundamental matrix with N rows and M columns indicating relatedness with each of first N objects and each of second M objects into three matrices, and clusters at least the first objects or the second objects. The data analysis method includes acquiring the fundamental matrix having each element storing a value indicating the relatedness, setting K indicating a number of clusters of the first N objects and L indicating a number of clusters of the second M objects, decomposing the fundamental matrix into three matrices which are a first matrix, a second matrix, and a third matrix such that a product of the first matrix, the second matrix, and the third matrix approximates the fundamental matrix, and outputting at least one of clustering results of the first objects and the second objects.

Abstract: A portable display device comprises: a tilt detection unit detecting an angle of rotation, measured between a vertical plane component of a tilt vector of a predetermined reference line in the display and a gravitational vector, as a display tilt; a facial orientation specification unit recognizing a face of a person facing the display and specifying an upright orientation of the face; a terminal orientation determination unit establishing a plurality of angle ranges, and determining whether the display tilt newly detected by the tilt detection unit exceeds a given angle range and enters another angle range, the given angle range including the angle detected by the tilt detection unit upon updating a previous display orientation; and a control unit causing the facial orientation specification unit to begin the specification when the terminal orientation determination unit makes an affirmative determination, and updating the display orientation according to the specified upright orientation.

Abstract: An acquisition section acquires a plurality of records which is a collection of records each of which includes a prediction value predicted by a prediction model and group information, a sorting section sorts the records based on the prediction values, a calculation section selects records from the sorted records in order, calculates the number of selections indicating the number of selected records and the number of correct decisions indicating the number of decisions in which the group information included in the selected record is decided to belong to a group classified as a correct decision, and outputs both the numbers in association with each other, an image drawing section generates a graph created by plotting both the numbers in a coordinate system with the horizontal axis representing the number of selections and the vertical axis representing the number of correct decisions, and a display section displays the generated graph.

Abstract: An object is to provide an apparatus that accurately estimates an association between content and a user who performed an operation on the content. The apparatus includes: a first association information generation unit that generates first association information in which content information subjected to an operation and a user who performed the operation are associated; a learning unit that generates learnt information for associating a given user with content likely to be operated by the given user by using the first association information; an estimation unit that estimates a user by using unidentified content information for which the user cannot be identified and the learnt information; and a second association information generation unit that generates second association information in which the unidentified content information and the estimated user are associated.

Abstract: Provided is a classification device capable of classifying time series data of location information into groups in which continuity of time is secured. The classification device (100) includes a data storage unit (110) that stores time-of-day information and location information corresponding to the time-of-day information, a sorting unit (120) that sorts the location information in order of time using the time-of-day information, a hidden Markov model storage unit (130) storing therein a left-to-right hidden Markov model containing a plurality of states and in which a parameter is undetermined, a Markov model output unit (140) that determines the parameter by inputting the location information sorted in order of time into the hidden Markov model, and a classification unit (150) that classifies each location information in one of the plurality of states by inputting the location information sorted in order of time into the hidden Markov model containing the determined parameter.

Abstract: A content output device for outputting contents to be displayed to users, including: a storage unit storing a plurality of contents, each being associated with its attribute; a first output control unit for outputting a predetermined content among the contents to a screen; a related content identifying unit for identifying, among the contents stored in the storage unit, a set of related contents whose attributes are related to an attribute of the predetermined content; a user preference identifying unit for identifying, among the contents stored in the storage unit, contents that match the preference of the user; and a second output control unit for outputting, to the screen, one or more contents that belong to both the set of related contents and the contents identified by the user preference identifying unit, thereby outputting images that are most likely to be desired by thy user.

Abstract: A co-clustering apparatus that performs co-clustering processing on relational data to divide the relational data into cluster blocks, the apparatus including: a distribution tendency generating unit that generates a distribution tendency of statistic amounts of the cluster blocks in the entire relational data, each of the statistic amounts indicating a tendency of relations generated in the corresponding cluster block; a calculate calculating unit that calculates an importance degree for each of the cluster blocks based on the statistic amount of the cluster block and the distribution tendency generated by the distribution tendency generating unit, using a calculation method for changing a result of calculation of the importance degree according to the distribution tendency; and an output unit that outputs at least one piece of information indicating the cluster blocks and information indicating the importance degree calculated for the at least one of information by the calculating unit.

Abstract: An object is to provide an apparatus that accurately estimates an association between content and a user who performed an operation on the content. The apparatus includes: a first association information generation unit that generates first association information in which content information subjected to an operation and a user who performed the operation are associated; a learning unit that generates learnt information for associating a given user with content likely to be operated by the given user by using the first association information; an estimation unit that estimates a user by using unidentified content information for which the user cannot be identified and the learnt information; and a second association information generation unit that generates second association information in which the unidentified content information and the estimated user are associated.

Abstract: A portable display device comprises: a tilt detection unit detecting an angle of rotation, measured between a vertical plane component of a tilt vector of a predetermined reference line in the display and a gravitational vector, as a display tilt; a facial orientation specification unit recognizing a face of a person facing the display and specifying an upright orientation of the face; a terminal orientation determination unit establishing a plurality of angle ranges, and determining whether the display tilt newly detected by the tilt detection unit exceeds a given angle range and enters another angle range, the given angle range including the angle detected by the tilt detection unit upon updating a previous display orientation; and a control unit causing the facial orientation specification unit to begin the specification when the terminal orientation determination unit makes an affirmative determination, and updating the display orientation according to the specified upright orientation.

Abstract: Provided is a classification device capable of classifying time series data of location information into groups in which continuity of time is secured. The classification device (100) includes a data storage unit (110) that stores time-of-day information and location information corresponding to the time-of-day information, a sorting unit (120) that sorts the location information in order of time using the time-of-day information, a hidden Markov model storage unit (130) storing therein a left-to-right hidden Markov model containing a plurality of states and in which a parameter is undetermined, a Markov model output unit (140) that determines the parameter by inputting the location information sorted in order of time into the hidden Markov model, and a classification unit (150) that classifies each location information in one of the plurality of states by inputting the location information sorted in order of time into the hidden Markov model containing the determined parameter.

Abstract: An information classification device of the present invention includes a classifying unit (3) classifying two or more content items into groups, each including a capturing time information item and a capturing location information item. The number of the groups is indicated in the number-of-groups information. Based on the capturing location information items, the classifying unit (3) merges each of the content items into the groups to generate first groups. In the case where there is a second group found among the generated first groups and including content items of which temporal continuity is not ensured, based on the capturing time information items, the classifying unit (3) divides the second group into sub-groups each including a content item of which temporal continuity is ensured, and adjusts the distance between each of two or more of the content items included in the sub-groups so that the sub-groups become distant from each other.

Abstract: A content output device for outputting contents to be displayed to users, including: a storage unit storing a plurality of contents, each being associated with its attribute; a first output control unit for outputting a predetermined content among the contents to a screen; a related content identifying unit for identifying, among the contents stored in the storage unit, a set of related contents whose attributes are related to an attribute of the predetermined content; a user preference identifying unit for identifying, among the contents stored in the storage unit, contents that match the preference of the user; and a second output control unit for outputting, to the screen, one or more contents that belong to both the set of related contents and the contents identified by the user preference identifying unit, thereby outputting images that are most likely to be desired by thy user.

Abstract: A video recording apparatus that prevents screen malfunctions due to playback of incomplete frame image data, even when a power failure has occurred while recording frame image data received via a network to a recording medium, and incomplete frame image data has been recorded. The video recording apparatus includes a reception unit 1001 that receives management information attached to each piece of data, a first recording unit 1003 that records the received video data to the recording medium 1010, and a second recording unit 1004 that records position information of the video data to the recording medium. The video recording apparatus quickly specifies damaged data positions by referencing a file length of the video data file 1005, and a position information file 1006.