Abstract:
PROBLEM TO BE SOLVED: To provide an information processor for analyzing network activity, an analyzing system, the analyzing method of the network activity, and a program. SOLUTION: The information processor 122 includes: an application providing part 270 which accepts access to information from a user computer, and generates network activity; an activity history storage part 128 which stores the activity history of the network activity; an infobabble generation part 220 which generates an infobabble by extracting information including a feature parameter, and acquiring network activity for the extracted information, and registering at least one user in an infobabble by using the user relationship in a network to the extracted information from the user of a user link with the user computer as a node; and a network activity analyzing part 240 which analyzes the network activity of the user by reading the infobabble. COPYRIGHT: (C)2010,JPO&INPIT
Abstract:
PROBLEM TO BE SOLVED: To allow a user to search by designating, through simple operation, a keyword suitable for searching in meeting the user' demand. SOLUTION: In a search support server 30, a related word extraction unit 32 generates frequency information and co-occurrence information of keywords, a graph generation unit 35 generates coordinate information of a spring graph including the keywords as nodes, on the basis of the co-occurrence information, a cluster generation unit 37 groups the nodes into clusters and thereby generates cluster definition information, and a display information generation unit 39 generates display information of the spring graph. In addition, an operation determination unit 42 determines which operation is performed on the spring graph. Then, when a level change is instructed, the display information generation unit 39 generates display information of the spring graph after the level is changed. When a node change is instructed, a cluster re-generation unit 43 changes the cluster definition information and the frequency information. When a search query generation is instructed, a search query generation unit 44 generates a search query with a keyword of a selected cluster. COPYRIGHT: (C)2010,JPO&INPIT
Abstract:
PROBLEM TO BE SOLVED: To provide a scalable link prediction technology that can cope with the number of dozens to millions nodes. SOLUTION: At first, similarity matrices W Z , W Y , W X , are low-rank approximated by a technology such as incomplete Cholesky decomposition. Then, the eigenvalue decomposition of low-rank approximate matrices of the similarity matrices W Z , W Y , W X is performed. Schematically, low-rank approximation is the approximation of one matrix by a product of two rectangular matrices. Here, low-rank approximation facilitates the calculation of eigenvalue decomposition. In the next step, eigenvalues of obtained low-rank approximate matrices of W Z , W Y , W X are used to constitute normalized Laplacian L. Since the normalized Laplacian L is obtained in this manner, V~ Z , V~ Y , V~ X as matrices with respective eigenvectors of low-rank approximate matrices of W Z , W Y , W X arranged therein and L are used to favorably calculate the inverse matrix of the part of (σL+I). When the inverse matrix of (σL+I) is obtained, F can be calculated due to vec(F)=(σL+I) -1 vec(F * ). COPYRIGHT: (C)2011,JPO&INPIT
Abstract translation:要解决的问题:提供可以应对数十百万个节点数量的可伸缩链路预测技术。 解决方案:首先,相似度矩阵W Z SB>,Y SB>,W X SB>通过诸如 不完全Cholesky分解。 然后,执行相似矩阵W Z SB>,W Y SB>,W X SB>的低阶近似矩阵的特征值分解。 示意地,低阶近似是由两个矩形矩阵的乘积的一个矩阵的近似。 这里,低阶近似有助于特征值分解的计算。 在下一步骤中,使用获得的W Z SB>,W Y SB>,W X SB>的低等级近似矩阵的特征值来构成归一化拉普拉斯算子L 由于以这种方式获得归一化拉普拉斯算子L,所以V = Z SB>,V = Y SB>,V = X SB> 布置在其中的W Z SB>,W Y SB>,W X SB>的低等级近似矩阵用于有利地计算部分的逆矩阵 (σL+ I)。 当获得(σL+ I)的逆矩阵时,可以由于vec(F)=(σL+ I) -1 SP> vec(F * SP>)计算F 。 版权所有(C)2011,JPO&INPIT
Abstract:
PROBLEM TO BE SOLVED: To provide an educational material selection method, a computer program, and an educational material selection device for selecting an optimal educational material on the basis of the learning evaluation data of an examinee. SOLUTION: The educational material selection method includes: a step wherein a computer calculates the achievement level of examinee's academic ability range on the basis of the examinee's learning evaluation data and the difficulty level of the academic ability range; a step for calculating the level of importance of the academic ability range on the basis of an academic ability range related structure representing interdependence between the achievement and the academic ability range; a step for calculating the value of the educational material on the basis of the level of importance; and a step for performing calculation for selecting an educational material on the basis of the value of educational material. COPYRIGHT: (C)2009,JPO&INPIT
Abstract:
Problem: Genaueres Erkennen von Gemeinschaften in einem sozialen Medium. Mittel zur Lösung des Problems: Es wird ein Verfahren zum Bilden von Clustern einer Vielzahl von Benutzern eines sozialen Mediums unter Verwendung eines Computers vorgeschlagen, wobei jeder aus der Vielzahl von Benutzern Nachrichten sendet. Das Verfahren weist die Schritte auf: Entnehmen einer Vielzahl von Teilgemeinschaften aus einer Vielzahl von Benutzern auf der Grundlage der Beziehungen von gleichartigen Nachrichten; Berechnen eines ersten Ähnlichkeitsgrades zum Anzeigen der Ähnlichkeit der gleichartigen Teilgemeinschaften auf der Grundlage der Beziehung zwischen einem Benutzer, der zu einer Teilgemeinschaft aus der Vielzahl von Gemeinschaften gehört, und einem Benutzer, der zu der anderen Teilgemeinschaft aus der Vielzahl von Gemeinschaften gehört; Berechnen eines zweiten Ähnlichkeitsgrades zum Anzeigen der Ähnlichkeit gleichartiger Teilgemeinschaften auf der Grundlage von Wörtern innerhalb der Nachrichten, die durch Benutzer gesendet wurden, die zu beiden Teilgemeinschaften gehören, und unter der Bedingung, dass die erste Ähnlichkeit einen vorgegebenen ersten Schwellenwert überschreitet; und Erzeugen einer vereinten Gemeinschaft durch Zusammenfassen der gleichartigen Teilgemeinschaften unter der Bedingung, dass die zweite Ähnlichkeit einen vorgegebenen zweiten Schwellenwert überschreitet.
Abstract:
PROBLEM TO BE SOLVED: To provide a method for more efficiently determining an optimum policy compared to an existing calculation method when a Markov decision process has cyclicity, and a device and a computer program therefor.SOLUTION: Provided is a method for determining an optimum policy by using a Markov decision process in which T (T is a natural number) pieces of subspaces, that have at least one states, have cyclic structure, respectively. The method includes steps of: identifying subspaces which are parts of a state space; receiving selection of t-th (t is a natural number and t≤T) subspace among the identified subspaces; calculating a probability and an expected value in costs of reaching from one or more states in the selected t-th subspace to one or more states in the t-th subspace of a following cycle; and recursively calculating a value and an expected value in costs on the basis of the calculated probability and expected value in costs, in a sequential manner starting from the (t-1)th subspace.
Abstract:
PROBLEM TO BE SOLVED: To derive an optimum arrangement configuration for allocating one or more virtual machines to one or more physical machines.SOLUTION: An arrangement configuration control device 120 includes a prediction unit 126 that determines a predicted peak usage amount of physical resources in each time section for each cluster containing a plurality of virtual machines having the same function; a setting unit 128 that sets a constraint condition for assuring that, when a physical machine 110 has failed in a time section, a predicted total peak usage amount of the physical resources for other physical machines 110 does not exceed a physical resource quantity prepared for the physical machines 110 for each combination of physical machines 110 and time sections; and an arrangement configuration deriving unit 132 for deriving the arrangement configuration by calculating a solution of an optimization problem that minimizes, as an object function, a total amount of physical resources of the whole of the plurality of physical machines to which the virtual machines are allocated according to the constraint condition.
Abstract:
PROBLEM TO BE SOLVED: To provide technology for classifying a network node. SOLUTION: The information processor 126 includes: an action history obtaining unit 210 for extracting an access log specified as a spammer-reporting action from the access log and generating a spammer-reporting action history set; a related node obtaining unit 220 for generating a node set and a link set related to the spammer-reporting actions; an undirected graph generation unit 230 for generating an undirected graph from the node set and the link set by registering a set of links connecting each pair of nodes as an edge in association with its link weight value; and a max-cut computation unit 240 classifying the nodes constituting the undirected graph into two exclusive sets that do not commonly include any element so as to maximize an indicator value defined by links bridging the two sets. COPYRIGHT: (C)2010,JPO&INPIT
Abstract:
PROBLEM TO BE SOLVED: To provide an analysis system, an information processor, an activity analysis method and a program. SOLUTION: In this analysis system, the information processor for analyzing activities of an information generation source on a network includes: a keyword information storage part 214 extracting a keyword from information transmitted on the network, and registering it; an information propagation graph acquisition part 212 generating action log data in association with information having an attribute characterized by an attribute designation keyword from action log data registered in association with an action type on the network related to the information, a user ID for peculiarly identifying a user, and the information, registering the information as a node in association with editing or the generation on the network, and generating a directed graph of the node, sequentially linked by a directed link; and a characteristic user calculation part digitalizing the activities as a measure by which the node of the directed graph functions as the information generation source. COPYRIGHT: (C)2009,JPO&INPIT
Abstract:
PROBLEM TO BE SOLVED: To provide a method, device and computer program for efficiently identifying items having a high frequency of occurrence among items included in a large-volume text data stream.SOLUTION: Identification information for identifying an item and a count for the item are stored in a memory of a higher level, and only identification information is stored in a memory of a level lower than said higher level. When text data stream input is received: if identification information for an item included in a bucket resulted from division of the received text data stream input is stored in the higher-level memory, the count for the item is incremented; when stored in the lower-level memory, the identification information for the item is transferred with an initial count to the higher-level memory; and, when not stored in any level, the identification information for the item is newly stored with the initial count in the higher-level memory.