40 results about "Web mining" patented technology

An enterprise-wide web data mining system, computer program product, and method of operation thereof, that uses Internet based data sources, and which operates in an automated and cost effective manner. The enterprise web mining system comprises: a database coupled to a plurality of data sources, the database operable to store data collected from the data sources; a data mining engine coupled to the web server and the database, the data mining engine operable to generate a plurality of data mining models using the collected data; a server coupled to a network, the server operable to: receive a request for a prediction or recommendation over the network, generate a prediction or recommendation using the data mining models, and transmit the generated prediction or recommendation.

The invention relates to a traveling domain knowledge map construction method and system, and belongs to the field of Web mining and intelligent information processing. Travelling domain knowledge map construction tasks include an entity attribute knowledge expansion subtask and an entity attribute value fusion subtask. By adopting a mixed entity attribute knowledge expansion method, an entity attribute knowledge expansion algorithm based on a lexical field, supervised learning, pattern matching and search engine questions and answers is integrated. For the entity attribute knowledge expansion subtask, a multi-value attribute value fusion method based on source credibility, a fixed pattern single-value attribute value fusion method based on content credibility and a non-fixed pattern attribute value fusion method based on learning ranking are adopted. The structured travelling domain entity knowledge base is constructed, traveling domain entity attribute and attribute value knowledge is accurately expressed, the traveling domain knowledge obtaining efficiency is improved, and the wide application prospect is achieved.

The invention relates to a multi-granularity semantic chunk based entity attribute and attribute value extracting method, and belongs to the technical field of Web mining and information extraction. The method comprises the following steps that a corpus set is constructed and free text extraction is performed; a corpus is subjected to word segmentation, part-of-speech tagging and phrase recognition; the corpus is subjected to semantic role labeling; the corpus is subjected to dependency grammar analysis; the corpus is subjected to semantic dependency analysis; candidate entities, attributes and attribute value triads based on three granularities of words, phrases and semantic roles are extracted; the candidate entities, attributes and attribute value triads are corrected and subjected to error classification by means of a trained classifier. Compared with the prior art, the entities, attributes and attribute value triads based on three granularities of words, phrases and semantic roles are automatically extracted from a free text, the entity attribute and attribute value extraction accuracy and efficiency are improved, and the wide application prospect is achieved in the fields of theme detection, information retrieval, automatic abstracting, question and answer systems and the like.

The invention relates to a coarse-grained emotion analysis method based on a hierarchical BERT neural network, and belongs to the technical field of Web mining and intelligent information processing.The method comprises the following steps: corpus acquisition: acquiring a corpus of coarse-grained sentiment analysis; corpus preprocessing, wherein character cleaning, subordinate clause segmentationand subordinate clause vector construction are included; constructing sentence vectors: calculating the subordinate clause vectors by utilizing a bidirectional long and short term memory network, a multi-layer perceptron and an attention mechanism to generate sentence vectors; gradient coordination mechanism optimizing: introducing the gradient coordination mechanism to solve the problem of datatype imbalance in coarse-grained sentiment analysis; and carrying out coarse-grained sentiment analysis by adopting a hierarchical BERT neural network. Compared with the prior art, t the sentence vectors containing deep semantic information are constructed for the comment text through the hierarchical BERT neural network, the accuracy of coarse-grained emotion analysis tasks is improved, and the method has a wide application prospect in the fields of information recommendation, public opinion monitoring and the like.

The invention discloses a video recommendation method and system based on Web mining. The method comprises the steps that a data mining algorithm is applied in clicking behavior data when users watch videos through Web mining, a user interest model is built through a classification and regression tree, a traditional collaborative filtering algorithm is adopted to recommend an individualized video to the users, the defect that in a traditional recommendation system, the data sparsity is brought due to the fact that user comment information is little is overcome, the problem of recommendation cold start due to the fact that a new user or a new project has no scores is solved, the satisfaction degree of the users to watch the video is improved, the users having the same interest and hobbies generate a recommendation, and friend recommendation is achieved in the video recommendation system.

The invention discloses a method for mining the orientation of Web themes and supporting decisions. The method includes steps of S1, extracting and storing network information, acquiring information on the internet via a network mining technology, and storing results in a database and a local file system; S2, detecting and tracking viewpoint themes of the information, detecting and identifying interested viewpoint themes with complete semantic information by the aid of thematic comment data and keeping tracking and following the viewpoint themes; and S3, identifying the emotion orientation of the viewpoint themes, classifying the emotion orientation of hot topics of an enterprise, and mining the emotion orientation of the view themes. The method has the advantages that related business information is acquired from the internet, so that the tendency of the orientation of the themes related to the enterprise can be quickly and effectively mined from the massive network information, real-time business intelligence can be realized, and decision support service can be effectively provided for the enterprise.

Methods for creating deep web mining plans from dynamic content filters are described. Dynamic content filters allow for the creation of deep web mining plans that are able to be used even when the structure of documents including web pages and PDF files changes or to apply the same filters to different variants of the pages generated in deep web mining. By basing the dynamic filters on ontological and semantic information many common changes in web page structure, terminology and format can be made without preventing the extraction of data from these pages in deep web mining. Dynamic content filters may be created by persons without expertise in the creation of deep web mining data extraction plans.

The invention discloses a thinking-map-based e-learning resource recommendation method, which comprises the following steps of analyzing a learning behavior log of a learner accessing a knowledge-map-based e-learning system, calculating a learning time length threshold value, and preprocessing the learning behavior log; constructing a learning path network and a learning transaction graph set; mining a thinking map (a bubble map or a double bubble map) for a specified knowledge element set on the basis of the learning path network; feeding back a mining result to the learner to realize thinking-map-based e-learning knowledge recommendation. According to the method, cognitive-strategy-based map type knowledge recommendation service can be provided for an e-learner, and the learning efficiency of the e-learner is further improved.

The invention relates to a method for predicting the gender of a microblog user based on deep learning and belongs to the field of Web mining and intelligent information processing. The prediction method includes the steps of collecting microblog information; preprocessing a microblog text; constructing word vectors of microblog text words; using a convolutional neural network-based microblog textrepresentation method to construct feature vectors of microblog text sentences; using a long-term and short-term memory network model-based method for gender prediction or classification of the microblog user. The convolutional neural network-based microblog text representation method can achieve semantic modeling of the microblog text without the need to manually construct microblog text features. The long-term and short-term memory network-based microblog user gender prediction method can extract semantic sequence dependency features in the microblog text. The method for predicting the gender of the microblog user can accurately extract the microblog text features and improve the recognition performance of the gender of the microblog user, and has broad application prospects in the fields of information recommendation and product marketing.

The invention provides an ash haze predicting system and method based on a BP neural network. The system excavates a relation among collected air quality data by means of the BP neural network, predicts air quality and ash haze weather development tendency, and provides early warning for the ash haze weather, and is characterized by comprising a data acquisition module, a database, a data arranging module, an ash haze prediction server, a data preprocessing module, a neuron learning module, a BP neural network prediction module, a WEB server and a mobile terminal. The method comprises steps of: describing the ash haze predicting system as a MIMO self-learning prediction system, inputting acquired ash haze and air quality data, predicting possible development tendency of ash haze weather by means of the self-learning and adaptive capability of the neural network, and reducing a prediction error. The ash haze predicting system and method may predict the ash haze development by using existing ash haze observations, deeply excavates a complex relation among input data, and obtain an accurate prediction result.

A data mining method for quickly finding a utility pattern can find a utility pattern which not only has substantial statistical characteristics but also meets user expectations and user goals from massive data, having a wide application in network information search and knowledge discovery. Aiming at solving the present problems of high time overhead and space overhead of existing methods caused by adoption of a two-stage method which generates a candidate pattern, the present invention provides three innovative technologies. The first is data representation based on a sparse matrix and virtual projection, the second is a prefix growth strategy, a prefix growth tree and a tailoring method thereof, and the third is a depth-first dynamic search method. With the three innovative technologies, a novel mining method is designed which has a single stage, causes no candidate pattern, and enables mining the utility pattern. The time efficiency ratio of the data mining method is higher by one to three orders of magnitude than that of other three referential mining methods, and the memory usage is reduced by 40% to 90%. The present data mining method has a high performance and enables various applications such as massive Web mining, multimedia mining and test mining.

The present invention discloses a method, a system and a computer program product for spend data classification using selected taxonomies. The invention provides refresh classification tool and implementation classification tools for spend data classification. The invention further provides a web mining tool for determining unknown terms in spend data to obtain an enriched spend data for classification.

The invention discloses a multi-layer frequent pattern discovery algorithm with high space extensibility and high time efficiency for mining mass data, relating to the field of intelligent information processing, which has a wide application prospect in mass data mining, particularly in network information search and knowledge discovery. Aiming at the problem that time and space expense bottleneck exists during mass data mining due to the present multi-layer algorithm for simply extending a single-layer algorithm, the invention provides three new technologies: the first one is a hierarchical labeling technology capable of integrating hierarchical structure information in a plurality of data expression methods by the least additional expense to solve the space expense bottleneck; the second one is an extensive virtual projection method for avoiding the repeat generation of a pattern support set and having a higher space utilization rate; the third one is an inverted set enumeration tree for organizing multilayer patterns and a cutting technology thereof, and the inverted set enumeration tree greatly reduces the search space of a frequent pattern, thereby solving an operation time bottleneck. The time and efficiency of the algorithm disclosed by the invention are about 5 times and 1-3 orders of magnitude higher than those of two reference algorithms and the space expense is the least. Various applications such as mass Web mining, multimedia mining and text mining become possible due to the high performance of the algorithm disclosed by the invention.

A system and method are provided for augmenting information on business directory databases and communicating with businesses is disclosed. Using the enriched business directory database and Web mining technology, customized email message are sent inviting businesses to enter their enriched business information into the directory or even subscribe to other paid services provided by the directory service.

The invention discloses a source code author identification method based on a deep belief network, and belongs to the field of Web mining and information extraction. The method includes the followingsteps of constructing a source code data set, and preprocessing source code data; extracting source code features based on a continuous n-gram code segment model; training a deep belief network modelbased on a training source code file sample; using the trained deep belief network model to identify an author of an source code file, and outputting an author identification result of the source codefile. The method converts a source code author identification problem into a classification problem, and identifies the identity of the author of the source code through the deep belief network, so that the performance and efficiency of identification of author identity are improved, and the method has broad application prospects in the fields of information retrieval, information security, computer forensics and the like.

The invention provides an unnormalized language processing method base on web mining, which relates to the filed of computer data mining, in particular to the technology of the network emotion mining scheme. The invention discloses a method for processing the network unnormalized language, which belongs to the field of computer data mining. The method provides a method for processing the unnormalized language by using the minimized monitoring study. The types of the normal unnormalized language are simplified from six kinds into two disjoint kinds: the typical unnormalized language and the ambiguous unnormalized language. The invention provides a model matching algorithm based on the sequence coverage for the typical unnormalized language, and provides a classification algorithm based on the feather extraction for the ambiguous unnormalized language to process the ambiguous unnormalized language. Finally, the completely normalized written words can be obtained, so the subjective opinion type mining operation is convenient, and information such as motion, opinions, advices and the like can be perfectly extracted.

This invention relates to building a landmark database from web data. In one embodiment, a computer-implemented method builds a landmark database. Web data including a web page is received from one or more websites via one or more networks. The web data is interpreted using at least one processor to determine landmark data describing a landmark. At least a portion of the landmark data identifies a landmark. Finally, a visual model is generated using the landmark data. A computing device is able to recognize the landmark in an image based on the visual model.

The invention relates to the field of complex network mining, in particular to a method for estimating the number of complex network communities. The method for estimating the number of community in complex network is disclosed. The complex network is denoted as graph and denoted as network G (V, E). The network G (V, E) contain m network nodes, denoted as V= (v1, v2,..., vm), wherein the ith network node (1 <= i <= m) is denoted as vi; There are n connections between nodes in the network, which are represented by n edges, denoted as E= (e1, e2, ..., en), where the first edge (1 <=l<= n) is denoted as el; the number of communities contained in the network G (V, E) and the community center nodes are determined. Without any prior information, it can fully reflect the intrinsic structural characteristics of dense network community and sparse network externality. Estimation accuracy of the number of communities in the network is high, which is conducive to improve the estimation performance of network community discovery algorithm, and has high practical value for the analysis of real network data.

The invention relates to a DNN based deep bottleneck feature extraction method of a heart impact signal, and relates to the technical field of biological feature extraction. The heart impact signal serves as a feature extraction object, and aimed at the characteristics that the heart impact signal is low in waveform amplitude and easy to be interfered by the outside, a deep bottleneck feature parameter is extracted from the heart impact signal by combining an electrocardio signal synchronously and using the mechanism that DNN digs a deep feature. The feature takes the heart impact signal as aninput vector and the synchronous electrocardio signal as a target vector, training is carried out via the pre-designed 9-layer neural network to obtain the deep bottleneck feature, and cardiodynamicsperformance is effectively combined with an electrophysiological feature. The feature takes the heart impact signal and the electrocardio signal easy to obtain daily as the research object, rely on waveform fluctuation of a routine waveform feature parameter is overcome, the representing performance of the single feature parameter can be improved, and the method serves as a new trial in daily heart function analysis by using the deep learning theories.