Abstract: A learning method includes: counting any one of or some of the number of labels added to each of feature amount vectors included in a learning data set, the number of types of the label, the number of feature amount vectors added with the same label, and the number of data pairs used for learning of a hyperplane, by a processor; first selecting, according to a result of the counting, one or more generation methods from a plurality of previously stored generation methods that generate the data pairs from the learning data set, by the processor; generating, using the selected generation methods, the data pairs from the feature amount vectors included in the learning data set, by the processor; and first learning, using the generated data pairs, the hyperplane that divides a feature amount vector space, by the processor.

Abstract: A memory stores second trajectory data indicating a plurality of reference positions, and road data indicating which road each of the reference positions belongs to. A processor acquires first trajectory data associating a plurality of times with a plurality of measured positions of a moving object, and velocity data indicating a moving velocity of the moving object at each of the times. The processor detects a first measured position where the moving velocity is equal to or less than a threshold. The processor extracts two or more reference positions located in a predetermined range from the first measured position. The processor complements the first trajectory data with information on roads corresponding to the respective times, by determining which road is to be associated with a time of measurement of the first measured position, based on the two or more reference positions and the road data.

Abstract: An information conversion method includes: first moving positions of a plurality of particles on a unit sphere according to a value of a probability density function, defining a positional vector of a particle on the unit sphere in a multidimensional space, as a normal vector of a hyperplane configured to divide a feature vector space, defining a predetermined evaluation function configured to evaluate the hyperplane, as the probability density function configured to indicate a probability of existence of a particle on the unit sphere, by a processor; and converting the feature vector to a binary string, considering a positional vector of the moved particle as a normal vector of the hyperplane, by the processor.

Abstract: A data classifying device performs processing request to first data groups stored in a database, obtains a parameter obtained from results of the performed processing for each data included in the first data group, and extracts a second data group from at least the plurality of first data groups based on a first similarity between the parameters and generates a third data group by classifying data included in the second data group so that second similarities between the parameters of the data included in the second data group are low and classifies the third data group to the first data group so that a third similarity between the parameters of the data included in a pair of the third data group and the first data group is low.

Abstract: A data arrangement apparatus executes a process request on data stored in a database, calculates similarity of a process in a combination of the process request and the data, divides the data in such a manner as to distribute the similarity and processes a process request accepted for the divided data.

Abstract: A mobile terminal includes an obtaining unit that obtains position information and time information of a plurality of satellites; a first output unit that outputs a variable for use for projection of a first coordinate system, to a projection surface defined in a second coordinate system, the second coordinate system being higher in dimension than the first coordinate system representing the position information and time information; a second output unit that outputs coordinates of the second coordinate system where the projection surface is present, the coordinates being computed using the position information and time information and the variable; and a transformation unit that transforms the coordinates output by the second output unit to coordinates of the first coordinate system.

Abstract: A method comprises extracting a picture type and size per picture frame from encoded first moving picture data, calculating a first characteristic quantity per picture frame and a first variation point indicating a variation point of the first characteristic quantity on the basis of the extracted picture type and size, extracting a picture type and size per picture frame from encoded second moving picture data, calculating a second characteristic quantity per picture frame and a second variation point indicating a variation point of the second characteristic quantity on the basis of the extracted picture type and size, and comparing the first characteristic quantity and variation point with the second characteristic quantity and variation point, and determining whether or not the first moving picture is similar to the second moving picture.

Abstract: A data placement device creates a similarity index for each of computational resources based on a similarity between each of the pieces of acquired data and each of the pieces of data stored in the computational resources. The data placement device allocates on the basis of the similarity index of each of the computational resources with respect to the pieces of the data, the pieces of the data to each of the computational resources by using a matching system in which the similarity index associated with each allocation becomes stable in a direction in which the similarity index is small. The placement device places the pieces of the acquired data into the computational resources on the basis of the allocation result.

Abstract: An information conversion method includes: first moving positions of a plurality of particles on a unit sphere according to a value of a probability density function, defining a positional vector of a particle on the unit sphere in a multidimensional space, as a normal vector of a hyperplane configured to divide a feature vector space, defining a predetermined evaluation function configured to evaluate the hyperplane, as the probability density function configured to indicate a probability of existence of a particle on the unit sphere, by a processor; and converting the feature vector to a binary string, considering a positional vector of the moved particle as a normal vector of the hyperplane, by the processor.

Abstract: A learning method includes: counting any one of or some of the number of labels added to each of feature amount vectors included in a learning data set, the number of types of the label, the number of feature amount vectors added with the same label, and the number of data pairs used for learning of a hyperplane, by a processor; first selecting, according to a result of the counting, one or more generation methods from a plurality of previously stored generation methods that generate the data pairs from the learning data set, by the processor; generating, using the selected generation methods, the data pairs from the feature amount vectors included in the learning data set, by the processor; and first learning, using the generated data pairs, the hyperplane that divides a feature amount vector space, by the processor.

Abstract: A learning method includes: randomly selecting one or more feature vectors from feature vectors for learning to form a sample set, by a processor; selecting, from the feature vectors for learning, one of feature vectors appended with a label different from a label appended to a feature vector included in the sample set as a reference vector, the selecting being carried out based on a generalized average of distance from a feature vector included in the sample set, by the processor; and learning a hyperplane that divides a feature vector space, the learning being carried out using a pair of one of feature vectors appended with a label different from a label appended to the reference vector, among the feature vectors for learning, and the selected reference vector, by the processor.

Abstract: A storage unit stores information on a graph corresponding to a structure data set that represents a three-dimensional structure with a plurality of polygons. An operation unit obtains the information on the graph from the storage unit. The operation unit determines a degree and parameter to be used in the operation of a symmetric polynomial, on the basis of the information on the graph, and then calculates a feature value of the graph corresponding to the structure data set with the symmetric polynomial using the determined degree and parameter.

Abstract: A storage unit stores structure data that includes coordinates of vertices of a plurality of polygons representing a three-dimensional structure. A computation unit calculates coordinates of a certain point with reference to the structure data. This point is be used, together with the vertices of the polygons, to produce a graph from the structure data according to spatial arrangement of the polygons and further to calculate characteristic quantities based on the produced graph. The computation unit stores the calculated coordinates of the point in a memory device, as a piece of information relating to the structure data.

Abstract: A classification method includes: setting a position of a projective point based on information on a domain of definition of a feature value space, using a processor; and projecting a vector present in the feature value space to a sphere present in a space of dimensions higher by at least one dimension than a dimension of the feature value space by using the set position of the projective point, using the processor.

Abstract: An information conversion device includes a memory and a processor coupled to the memory. The processor executes a process including generating a conversion rule for converting a feature quantity vector into a binary string that is longer than a predetermined bit length. The process includes converting each of the feature quantity vectors into a binary string by using the conversion rule generated at the generating. The process includes calculating importance levels of the respective bits in the binary strings based on the distance-based relations among the feature quantity vectors. The process includes correcting the conversion rule into one that converts each of the feature quantity vectors into a binary string having the predetermined bit length based on the calculated importance levels. The process includes changing the feature quantity vectors into a binary string having the predetermined bit length by using the conversion rule corrected at the correcting.

Abstract: A search device according to an embodiment maps a feature vector onto a hyper-sphere on the basis of parameters which include an intersection and a distance, with the intersection at which an m-dimensional feature space and a straight line passing through the hyper-sphere present in a space greater in dimension than m intersect and the distance being from the north pole of the hyper-sphere to the feature space. In this case, the search device searches for the parameters which allow the positions of feature vectors mapped onto the hyper-sphere to be concentrated on a predetermined hemisphere of the hyper-sphere.

Abstract: A space division method includes specifying and placing. The specifying includes specifying a position of an intersection of hyper-planes such that the position is contained within a sphere present in a space of a dimension higher than a dimension of a feature space by one dimension or more, by a processor. The placing includes placing the hyper-planes so that the hyper-planes share the intersection of the specified position, by the processor.

Abstract: With a search device according to the present embodiment, where an inverse stereographic projection unit performs an inverse stereographic projection of a feature vector onto a hyper-sphere divided into areas by a plurality of hyper-planes, and a bit string generating unit specifies a bit string on the basis of the results, a placement unit sets the initial positions of the hyper-planes. For example, the placement unit sets the initial positions of the hyper-planes so that the intersections of the hyper-planes are contained within the hyper-sphere.

Abstract: An evaluation server includes a memory and a processor coupled to the memory, wherein the processor executes a process includes creating, from a plurality of pieces of data in a database, matching data that is to be matched with the plurality of pieces of data, calculating matching time for each of the plurality of pieces of data in the database, the matching time being time taken to match each of the plurality of pieces of data with the matching data, and allocating the plurality of pieces of the data to a plurality of groups based on the matching time for each of the plurality of pieces of data.

Abstract: A method comprises extracting a picture type and size per picture frame from encoded first moving picture data, calculating a first characteristic quantity per picture frame and a first variation point indicating a variation point of the first characteristic quantity on the basis of the extracted picture type and size, extracting a picture type and size per picture frame from encoded second moving picture data, calculating a second characteristic quantity per picture frame and a second variation point indicating a variation point of the second characteristic quantity on the basis of the extracted picture type and size, and comparing the first characteristic quantity and variation point with the second characteristic quantity and variation point, and determining whether or not the first moving picture is similar to the second moving picture.