Abstract: Embodiments of the present invention are directed to methods and systems for determining unknowns in rating matrices. In one embodiment, a method comprises forming a rating matrix, where each matrix element corresponds to a known favorable user rating associated with an item or an unknown user rating associated with an item. The method includes determining a weight matrix configured to assign a weight value to each of the unknown matrix elements, and sampling the rating matrix to generate an ensemble of training matrices. Weighted maximum-margin matrix factorization is applied to each training matrix to obtain corresponding sub-rating matrix, the weights based on the weight matrix. The sub-rating matrices are combined to obtain an approximate rating matrix that can be used to recommend items to users based on the rank ordering of the corresponding matrix elements.

Abstract: Systems, methods, and machine readable and executable instructions are provided for collaborative filtering. Collaborative filtering includes representing users and objects by rows and columns in a binary ratings matrix having a particular dimensional space. Unknown values in the binary ratings matrix are weighted with a weight matrix having the particular dimensional space. The binary ratings matrix and the weight matrix are hashed into a lower dimensional space by one of row and column. The hashed binary ratings matrix and the hashed weight matrix are low-rank approximated by alternating least squares. A result of the low-rank approximation for the one of row and column is updated using the binary ratings matrix and the weight matrix. A recommendation of one of the objects can be generated for one of the users based on the updated result.

Abstract: A method for data clustering may comprise entering data into a computer network comprising a master processor, an array of slave processors, and two cluster seats associated with each slave processor; executing a master process comprising dividing the data into clusters, sending the clusters to the cluster seats, initializing an optimization cycle, and computing an objective function. The optimization cycle includes the parallel execution by the slave processors of a slave process, which includes exchanging data between paired clusters so as to increase the objective function based on two modalities, and then resorting the cluster pairs for a subsequent iteration of the process.

Abstract: For each web page visited, a path is determined through a hierarchy of categories. The hierarchy of categories has levels from a most abstract level to a most concrete level. For each microblog entry of a microblog, a path is determined through the hierarchy of categories. Each microblog entry for which the path is similar to the path for at least one web page is determined as a selected microblog entry.

Abstract: A collaborative filtering method for evaluating a group of items to aid in predicting utility of items for a particular user comprises assigning an item value of either known or missing to each item of the group of items, and applying a modification scheme to the item values of the missing items to assign a confidence value to each of the item values of the missing items to thereby generate a group of modified item values. The group of items having modified item values and the group known items are evaluated to generate a prediction of utility of items for a particular user.

Abstract: For each first entity of a subset of a number of first entities, an expected improvement of a predictive performance of a collaborative filtering model if additional ratings of the first entity in relation to a plurality of second entities were obtained is estimated. Particular first entities from the subset of the first entities of which to obtain the additional ratings in relation to the second entities are selected based at least on the expected improvements that have been determined. The additional ratings of the particular first entities in relation to the second entities are obtained.

Abstract: An exemplary embodiment of the present invention provides a method of generating Website content. The method includes generating a client profile comprising a cluster type obtained from a list of cluster types and information received from a user ID, wherein the list of cluster types is generated by processing a database of computer usage. The method includes utilizing the relevant cluster types included in the client profile to a selected Website, wherein the cluster type is used by the Website at least in part to determine the content provided by the Website.

Abstract: One embodiment is a method that receives a seed Uniform Resource Locator (URL) that represents a category for website classification. Clickstream data generated from the seed URL and additional URLs are analyzed to determine whether the additional URLs belong to the category. The method selects one or more of the additional URLs to represent the category.

Abstract: Embodiments of the present invention are directed to methods and systems for determining unknowns in rating matrices. In one embodiment, a method comprises forming a rating matrix, where each matrix element corresponds to a known favorable user rating associated with an item or an unknown user rating associated with an item. The method includes determining a weight matrix configured to assign a weight value to each of the unknown matrix elements, and sampling the rating matrix to generate an ensemble of training matrices. Weighted maximum-margin matrix factorization is applied to each training matrix to obtain corresponding sub-rating matrix, the weights based on the weight matrix. The sub-rating matrices are combined to obtain an approximate rating matrix that can be used to recommend items to users based on the rank ordering of the corresponding matrix elements.

Abstract: An exemplary embodiment of the present invention provides a method of processing Web activity data. The method includes obtaining a database of Website organizational data. The method also includes generating a data structure from the database of Website organizational data comprising an Item identifier and a Website category corresponding to the item identifier. The method also includes generating a reduced-rank classification structure from the data structure, the reduced-rank classification structure including a category grouping corresponding to one or more of the Website categories.

Abstract: An exemplary embodiment of the present invention provides a method of receiving Website content. The method includes generating a user profile comprising a cluster type obtained from a list of cluster types, wherein the list of cluster types is generated by processing a database of search queries. The method includes providing the relevant cluster types included in the user profile to a selected Website, wherein the cluster type sent to the Website is used by the Website at least in part to determine the content provided by the Website.

Abstract: A computer-implemented method comprises receiving, by a processor, a vector and a matrix. The vector includes historical periodic returns of a portfolio and the matrix contains historical periodic returns of each security in a set of all possible securities comprising the portfolio. The method further comprises computing, by a processor, a linear programming solution of a vector of weights of the securities comprising the portfolio. The vector comprises a product of the matrix and the vector of weights. The linear programming solution is subject to a criterion that a sum of absolute values of the weights in the vector of weights is a minimum. The method also comprises displaying or storing results of the computing.

Abstract: An exemplary embodiment of the present invention provides a method of processing Web activity data. The method includes obtaining a database of clickstream data comprising a user identifier corresponding with a user ID and a uniform resource locator (URL) corresponding with a Web page visited from the user ID. The method also includes generating a plurality of features based on the URL. Further, the method includes generating a data structure comprising the user ID and the feature. The method also includes generating segment information from the data structure based on the similarity of a URL visitation pattern across different user IDs, wherein each segment in the segment information comprises one or more user IDs and one or more features.

Abstract: A collection of labeled training cases is received, where each of the labeled training cases has at least one original feature and a label with respect to at least one class. Non-linear transformation of values of the original feature in the training cases is applied to produce transformed feature values that are more linearly related to the class than the original feature values. The non-linear transformation is based on computing probabilities of the training cases that are positive with respect to the at least one class. The transformed feature values are used to train a classifier.

Abstract: A system and method for providing personalized recommendations are disclosed herein. A system includes a processor and a software system executed by the processor. The software system provides a recommendation for an item. The recommendation is based on a comparison of a low-rank approximation of a domain matrix to a user profile. The user profile is based, in part, on the low-rank approximation of the domain matrix.

Abstract: An exemplary embodiment of the present invention provides a computer implemented method for classifying information. The method may include accessing a plurality of information sources to identify example information items for each of a plurality of classification categories. Each of the example information items may be analyzed to generate a training corpus for each information source for each of the classification categories. The training corpus for each of the information sources may be combined to generate a training set for each of the classification categories, wherein the training set may be configured to allow the generation of a classification function.

Abstract: A computer-implemented method comprises running, by a processor, a plurality of classifiers on a web page to obtain one or more keywords. The method further comprises selecting, by the processor, advertisements based on the one or more keywords.

Abstract: There is described a system and computer-implemented method for providing a recommendation based on a sparse pattern of data. An exemplary method comprises determining a likelihood that an item for which no user preference data is available will be preferred. The exemplary method also comprises determining a likelihood that an item for which user preference data is available for users other than a particular user will be preferred based on the likelihood that the item for which no user preference data is available will be preferred. The exemplary method additionally comprises predicting that an item for which no user preference data relative to the particular user is available will be preferred if the likelihood that the particular user will prefer the item exceeds a certain level.

Abstract: A collaborative filtering method for evaluating a group of items to aid in predicting utility of items for a particular user comprises assigning an item value of either known or missing to each item of the group of items, and applying a modification scheme to the item values of the missing items to assign a confidence value to each of the item values of the missing items to thereby generate a group of modified item values. The group of items having modified item values and the group known items are evaluated to generate a prediction of utility of items for a particular user.

Abstract: A method of classifying a plurality of documents that form part of a data set comprises retrieving the plurality of documents from a computing device and applying a hashing representation scheme to the plurality of documents from the data set to obtain a feature vector representation of each of the plurality of documents. A classification label is associated with selected documents of the plurality of documents in the data set. A learning algorithm is executed to learn a functional relationship between the feature vector representations of the plurality of documents and the classification label associated with the at least one document. The functional relationship learned is utilized to associate classification labels with feature vector representations of other documents of the data set so as to provide document classifications.