867 results about "Decision tree model" patented technology

A text categorization method automatically classifies electronic documents by developing a single pooled dictionary of words for a sample set of documents, and then generating a decision tree model, based on the pooled dictionary, for classifying new documents. Adaptive resampling techniques are applied to improve the accuracy of the decision tree model.

The invention relates to a stratification characteristic analysis-based method and an apparatus thereof for on-line identification for TCP, UDP flows. The method comprises the following steps that: an off-line phase determines a common port number of a first layer to-be-identified service type and a characteristic field of a second layer to-be-identified service data flow through a protocol analysis; a port number and characteristic field database is constructed; meanwhile, a third layer Bayesian decision tree model is obtained by training by employing a machine study method; and service type identification on a flow is completed by utilizing the characteristic database and a study model at an on-line classification phase. In addition, the apparatus provided in the invention comprises a data flow separating module, a characteristic extraction module, a characteristic storage module, a characteristic matching module, an attribute extraction module, a model construction and classification module and a classification result display module. According to the embodiment of the invention, various application layer services based on TCP and UDP are accurately identified; moreover, the identification process is simple and highly efficient; therefore, the method and the apparatus are suitable for realization of a hardware apparatus and can be applied for equipment and systems that require on-line flow identification in a high speed backbone network and an access network.

The invention relates to an individualized singing sound synthesis method, including the following steps: building up a module of the coefficient of the line spectrum frequency of the sound and obtaining a relevant decision-making tree module through training; recording the reading sounds of a special subscriber to get the module of the coefficient of the line spectrum frequency of the sound of the subscriber; obtaining the attribute set relevant to the context of the lyric of the numerical notes, and pre-estimating the frequency parameters and the time duration data of initial consonant and vowels corresponding to the lyric according to the decision-making tree module and the module of the coefficient of the individualized line spectrum frequency; building up fundamental frequency data according to the numerical notes and combining the fundamental frequency data with the time duration and frequency parameters to obtain synthesized parameters; inputting the parameters into a parameterized sound synthesis vocoder, so that individualized singing sound can be synthesized. The method of the invention can synthesize synthesis sound with singing style by adjusting a few parameters of the rhythm and can synthesize singing sound by only recording a small reciting style library.

The invention discloses a shopping behavior prediction method, and the method comprises the steps: selecting different target users at different shopping stages, and obtaining sample data from a user behavior log of the selected target users; respectively extracting a first feature set for marking user behavior from the sample data at each shopping stage; respectively training the first feature sets at different stages through a decision tree model, obtaining a feature combination through multiple iteration, and enabling the feature combination to serve as a second feature set; respectively employing the first and second feature sets at all shopping stages to train a pre-built machine learning model, wherein the machine learning model is used for predicting the shopping demand degree of a user; and determining the shopping stage of a to-be-detected user according to the machine learning model at different shopping stages. The invention also provides a corresponding shopping behavior prediction device.

The invention discloses a coastal city time sequence land utilization information extracting method. The method comprises the following steps: acquiring a remote-sensing image Landsat, and preforming atmospheric correction on the same; constructing a remote-sensing classification feature index database by selecting a group of remote-sensing classification features; acquiring data elevation image DEM data to obtain elevation data and slope data; constructing a decision rule of single-classification feature index or multiple classification feature indexes according to different land utilization types of the coastal city based on a multi-feature decision tree model, classifying the coastal city land utilization step by step according to the rule, and finally determining various branches of the decision tree, detecting the time sequence remote-sensing image change, and distinguishing a mistaken classification land type and a missed classification land type, wherein the method further comprises the content of two parts: evaluating classification precision, and outputting the land utilizationclassification map extracted based on the decision tree model. By use of the extracting method disclosed by the invention, the coastal city land utilizationclassification precision can be greatly improved, and a key problem in the coastal city land utilizationclassification is solved.

The invention discloses a vehicle forced lane changing decision-making method based on a decision-making tree model. The vehicle forced lane changing decision-making method includes the following steps: firstly, reading related data during vehicle forced parallel lane changing in real time through a sensor; secondly, importing the obtained data into a vehicle forced lane changing decision-making module based on the decision-making tree model, wherein a method for building the module includes the steps of selecting training and testing data, splitting a tree, selecting attribute threshold values, pruning the tree, building the parallel lane changing decision-making tree model based on a weka platform and verifying the accuracy of the decision-making model; finally, forming a decision-making judgment result during vehicle forced lane changing through a decision-making module, and if the decision-making judgment result is that lane changing can not be carried out, giving an alarm in real time to remind a driver of the fact that lane changing can not be carried out. By means of the vehicle forced lane changing decision-making method, negative effects, caused by a complex early-warning algorithm and excessive decision-making judgment rules, on the judgment result are reduced, the accuracy and the reliability of decision-making judgment during vehicle forced lane changing are improved, and the false alarm rate is lowered.

The invention discloses a state prediction method and device. The method comprises the following steps of: sampling target users; respectively generating a negative sample, a positive sample and a verification sample according to account information of lost and unlost users according to a recognized sampling moment; training a decision-making tree which is used for predicting user loss state afterthe sampling moment; inputting the verification sample into the trained decision-making tree model to obtain a predicted loss state; and if a correct recall rate obtained through carrying out calculation according to the predicted loss state and a practical loss state is not smaller than a threshold value, determining that the training of the decision-making tree model is completed, and carryingout user loss state prediction. According to the state prediction method and device, the decision-making tree model is trained by a training sample generated through sampling the target users, and user loss state prediction is carried out according to the trained decision-making tree model, so that the technical problems of relatively recognition efficiency low and reusing difficulty caused by user loss state prediction carried out through artificial experiences or established rules in the prior art are solved.

The invention discloses a travel time prediction method based on multi-modal data fusion and multi-model integration. The travel time prediction method comprises a multi-modal data preprocessing module which extracts taxi travel data from taxi GPS track data according to the passenger carrying state; a multi-modal data analysis, feature extraction and feature fusion module which is used for extracting corresponding feature sub-vectors from the fields of taxi track data, weather data, driver portrait data and the like and completing feature splicing; and a multi-model integration module which is used for respectively establishing a gradient improvement decision tree model and a deep neural network model, and integrating prediction results of the models by using the decision tree model. According to the travel time prediction method, by fusing the multi-modal data such as taxi track data, weather data and driver portrait data, the factors influencing travel time are fully extracted and mined, and an integrated model based on a decision tree is established, so that higher travel time prediction accuracy is obtained at lower calculation cost.

Decision trees are efficiently represented in a relational database. A computer-implemented method of representing a decision tree model in relational form comprises providing a directed acyclic graph comprising a plurality of nodes and a plurality of links, each link connecting a plurality of nodes, encoding a tree structure by including in each node a parent-child relationship of the node with other nodes, encoding in each node information relating to a split represented by the node, the split information including a splitting predictor and a split value, and encoding in each node a target histogram.

An exemplary method and system for evaluating media-playing sets evaluates the likelihood of a selected performance condition occurring in a subject set including based on source data automatically collected from a sample group of the sets, systematic analysis of this data to form a decision tree model revealing prescribed values for characteristic input parameters that are determined to best relate to the condition, and automated comparison of the respective parameter values of the subject set to these prescribed values in order to screen each subject set for the likelihood of the condition occurring within a specified timeframe, which screening can be repeated for different conditions and timeframes using different decision tree models.

The present invention provides a method and a system for optimizing classification of a random forest based on weighted decision trees, belonging to the mode identification technology field. The method comprises the steps of: employing a bootstrap method to generate a plurality of training data sets; randomly extracting one feature set for each training data set; training decision trees, and distributing a voting weight of each decision trees according to statistic features of the feature sets or performances of the decision trees; and introducing a voting mechanism, and accelerating a classification process of a random forest. The method and the system for optimizing classification of a random forest based on weighted decision trees employ the statistic features of the feature sets or theperformances of the decision trees to distribute voting weights of the decision trees, and employ the voting mechanism to accelerate the decision process so that the classification performances and the classification efficiency of the random forest are effectively improved.

The invention discloses a fuzzy rough set and decision tree-based track circuit red light strip fault positioning method. The method mainly comprises the following steps of: 1) establishing an initial decision table; 2) carrying out fuzzy discretization on continuous fault feature attributes to establish a fuzzy decision table; 3) inputting fault sample training data to obtain a reduced decision table; 4) establishing a diagnosis decision tree model; 5) inputting measured data into the diagnosis decision tree model, carrying out calculation to obtain a fault diagnosis result, inputting the measured data into a diagnosis positioning decision tree model, carrying out preliminary judgement to obtain a fault positioning result, judging faults of specific equipment by combining expert experiences, and giving corresponding fault maintenance suggestions. The method can rapidly and correctly position fault points of uninsulated frequency shift track circuit red light strip faults, greatly reduce the blindness and complexity of fault diagnosis, have relatively good rule explanation and relatively good robustness, improve the fault positioning speed and correctness and provide a new fault positioning technological means for intelligent fault diagnosis of track circuits.

The present invention discloses an order answering willingness evaluation method and a device for drivers in an online taxi service platform. The method comprises the steps of obtaining training samples, wherein each training sample comprises the order information, the driver state information and the information of whether a driver answers an order or not; obtaining a decision tree model through the training of training samples; according to the decision tree model, converting the order information and the driver state information of each of the training samples into attribute vectors so as to obtain converted training samples; according to the converted training samples, training to obtain a driver order-answering probability prediction model; according to the driver order-answering probability prediction model, determining the order-answering willingness of a driver. According to the technical scheme of the invention, the accuracy of evaluation results is improved.

The invention relates to an assessment method for transient voltage stabilization of a load area of an electrical power system and belongs to the electrical power system stability analyzing and assessment field. The assessment method for the transient voltage stabilization of the load area of the electrical power system includes that using data measured by a synchronous phasor measurement unit as basis, and building an initial sample database for data mining through a lot of simulation samples; extracting characteristics for reflecting the stabilization degree of each node through the quantitative evaluation for each node in the area; identifying through a multiple linear regression method to obtain sensitivity coefficients for reflecting the mutual influence relations between the nodes in a local area network; using a semi-supervised clustering method to demarcate all the samples; using a decision tree algorithm to perform classified learning to obtain a decision tree model, and using the decision tree model for online monitoring to assess the global transient voltage stabilization state of the load area of the electrical power system.

The invention provides a fast partitioning method of HEVC (High Efficiency Video Coding) intraframe coding units based on a decision tree. The method comprises the following steps: acquiring texture features of a current coding unit by constructing a histogram, wherein the texture features comprise the number of edge points of the current coding unit, the variance of a brightness value, the variance of a mean value of a subblock brightness value of the current coding unit and the variance of a subblock brightness value variance; and predicting the coding depth of the current coding unit by adopting a decision tree model according to the texture features. Since the texture features of the current coding unit are analyzed by constructing the histogram, and adaptive selection of the coding scale is performed by using the texture features through the decision tree model, thus unnecessary coding scale calculation can be skipped; and besides, experimental results show that the method provided by the invention saves 31% coding time on average during intraframe coding, while the average bit stream is increases by about 2.6% and at the same time, the PSNR of a video basically remains unchanged, thereby greatly reducing the coding complexity of the HEVC.

The invention discloses a neural network-based intrusion detection method, comprising the following steps: step 1), a detecting step which comprises the following steps: sniffing to capture traffic information of a connected host, determining which ports are open or closed and which programs are running, and judging whether the system has received an attack or is about to receive an attack according to the information; step 2), a data preprocessing step which comprises the following steps: giving a large number of training samples, selecting features, performing data preprocessing, and identifying anomalies; step 3), an attack classification step which comprises the following steps: using a decision tree model, a support vector machine or a neural network model to classify various attacksby using a neural network algorithm; and step 4), an alerting step which comprises the following steps: notifying the detected attack event, so that a network administrator can make decisions in timeand the damage caused by attacks is reduced.

The present invention belongs to the photovoltaic power generation technical field and provides a decision tree model-based photovoltaic assembly fault diagnosis method. The method includes the following steps that: photovoltaic assembly data are acquired, and data processing is carried out; obtained data are introduced into a decision tree-based photovoltaic assembly fault diagnosis model, the modeling steps of the model mainly comprise selection and processing of training and test sample data, tree establishment and tree pruning, decision tree model establishment and decision tree accuracy verification; and the fault type of a photovoltaic assembly is judged through a decision module. With the method of the invention adopted, manpower and resource waste caused by fault judgment errors can be avoided, the accuracy and reliability of fault judgment are improved, serious consequences to the photovoltaic assembly, caused by a fault, can be avoided, and the service life of the photovoltaic assembly can be prolonged.

The invention relates to an objectionable image distinguishing method integrating skin color, face and sensitive position detection. The method comprises the following steps that a skin color model is firstly built, the face detection is carried out, the constituted feature vector of skin color and face features is extracted, a SVM (support vector machine) algorithm is utilized for training, and a SVM classifier is obtained; then, by aiming at the female breast in the local key position of the human body, SIFT (scale-invariant feature transform) features are extracted, an Adaboost algorithm is utilized for training, and an Adaboost classifier is obtained; next, by aiming at the female private parts in the local key position of the human body, the trunk region of the human body is determined, haar-like features are utilized as a template for carrying out searching and matching in the trunk region of the human body; and finally, the SVM classifier, the Adaboost classifier and the template matching method are adopted for carrying out image detection, a C4.5 decision-making tree method is utilized for integrating detection results, a decision-making tree model is built, the decision-making tree model is adopted for recognizing objectionable images, and the final distinguishing results are given. The objectionable image distinguishing method has the advantages that the detection accuracy is improved, and meanwhile, the execution speed is ensured.

An embodiment of the invention discloses a method for releasing advertisements, a terminal and a computer readable storage medium. The method includes receiving total inquiry traffic and acquiring model training data and static attribute feature data and dynamic attribute feature data of users according to the total inquiry traffic; determining total scene data about exposure / clicking according tothe model training data, the static attribute feature data and the dynamic attribute feature data; combining logic regression models and gradient enhancement decision tree models with one another according to single-order features to predict a plurality of potential click-through rate values of the multiple advertisements; receiving advertisement click-through value estimated values inputted by users and computing offered prices of the multiple advertisements according to the multiple potential click-through rate values and the advertisement click-through value estimated values; ranking the offered prices of the multiple advertisements to obtain advertisement offered price ranking tables and releasing the advertisements according to the advertisement offered price ranking tables. The method, the terminal and the computer readable storage medium in the embodiment of the invention have the advantages that the pCTR (predict click-through rates) of advertisement click-through rates can beaccurately estimated, and the advertisements further can be accurately released.

The invention discloses a decision tree based game cheat detection method invented for solving the problem of the prior art that the passive defensive mode cannot detect external hanging. The method of the invention comprises the following procedures: the non-redundant characteristic attribute data set is extracted from the information database of a player at predetermined time; data in the characteristic attribute data set are divided into characteristic attribute exercise data and characteristic attribute test data; a decision tree is generated with the characteristic attribute exercise data and clipping on the decision tree through the characteristic attribute test data is performed to generate an objective decision tree model; the objective decision tree is assessed to obtain an appropriate objective decision tree model; analytical treatment is performed on the objective decision tree to produce categorized player databases and then on-line analysis of the categorized player databases is conducted to detect cheating players with external hanging. The invention can conduct data mining analysis with the data mining method in an active defensive mode to detect cheating players, thereby improving the external hanging preventing ability in online network games.

The invention belongs to the field of the Internet and provides a method and a device for identifying webpage categories. The method comprises obtaining page characteristics of a webpage to be identified, loading the page characteristics according to a pre-generated decision-making tree model which is generated by training a plurality of sample webpages of a determined webpage category, recursively traversing the decision-making tree model, searching for decision-making tree leaf classification nodes corresponding to the page characteristics, and obtaining the webpage category of the webpage to be identified according to the leaf nodes. The page characteristics of the webpage to be identified are obtained, the obtained page characteristics are loaded to the pre-generated decision-making tree model, and the decision-making tree model is generated by training the plurality of sample webpages of the determined webpage category, and therefore, the webpage category corresponding to the decision-making tree leaf classification nodes can be found quickly and effectively, and the extraction of webpage contents, the analysis of user behaviors and better display of the contents in the page in a cellphone browser can be realized conveniently.

The invention discloses a server fault automatic detection system and a detection method based on a decision tree, which combines an expert system and an IPMI management unit to generate a historicaldata set. The server running state data, i.e. Abnormal data stream, is obtained by IPMI management unit. According to the abnormal data stream, the new fault feature vector is extracted, and the new feature vector and the fault cause relation pair are formed into a fault data set, and the fault data set is trained into a self-diagnosing decision tree model. When the server fails while running, thecorresponding fault feature vectors are extracted, the fault types are automatically judged by the self-diagnostic decision tree model, After the fault is cleared, the fault feature vector and the fault cause relation are updated and the self-diagnosing fault tree model is updated. Therefore, the fault diagnosis system will be more accurate and reliable with the improvement of the history fault set.

An activity recognition method, for recognizing continuous activities of several moving objects in the foreground of a video, includes: capturing and processing a training video to get a contour of a moving object; extracting a minimum bounding box of the contour in order to get parameters then transfer to feature vectors; constructing a decision tree model based on support vector machines (SVMs), for classifying the activities of the moving object according to the parameter and the feature vector of the training video; capturing and processing a testing video to get other parameters and using several formulas to generate feature vectors, and executing an algorithm for recognizing the activities of several moving objects in the foreground of the testing video. Said feature vectors are transformed from the parameters that in the testing and training videos. Via above descriptions, the method can recognize activities of foreground objects in the testing video.

The invention provides a method and a system for processing a call based on a decision tree model; the method comprises the following steps: obtaining basic information and associated information of a user according to an incoming call number; combining the basic information and the associated information as an input and matched first pushing information as an output to construct the decision tree model; inputting the basic information and the associated information of the specific user to the trained decision tree model and calculating to obtain second pushing information of the specific user; and pushing the second pushing information to a customer service. According to the method and the system for processing the call based on the decision tree model provided by the invention, when the incoming call user dials a customer service hotline, the user does not need to input a digital incoming line again according to a voice prompt, and the call dialing is accurately matched to a customer service group, so that the artificial customer service, when the user incomes the line, can know the question that the user wants to consult simultaneously; and therefore, the operation of the user and the time of listening to a prompt tone for operation are effectively reduced, the user satisfaction of a customer is effectively improved, and the working efficiency of the customer service is improved.

The invention discloses a decision-making tree based prediction method and device and relates to the field of data processing, and the method and the device can improve the visualization effect of a decision-making tree model and the decision-making tree model testing process. The decision-making tree based prediction method includes steps of generating a target training set by feature selection of data in a defective data set of prestored products according to feature attributes; training a training set by a decision-making tree algorithm to generate a decision-making tree, wherein the training set belongs to the target training set; compressing the decision-making tree to obtain a first decision-making tree; displaying the first decision-making tree by means of the visualization technology; selecting at least one test case from a test set to be sequentially input into the first decision-making tree for testing, and generating classification paths for the test cases; displaying the classification paths of the test cases in the first decision-making tree by means of the visualization technology. The decision-making tree based prediction method and device is applied to defect prediction of the products.

The invention discloses a user behavior prediction method and device, relates to the technical field of information and is mainly intended to predict user behaviors more accurately. The method comprises: analyzing a user behavior log of a target user to obtain a user attribute vector and a user behavior index vector; inputting the user attribute vector and the user behavior index vector to a pre-trained behavior prediction decision tree model to obtain predicted behavior probability of the target user, wherein the behavior prediction decision tree model is acquired by using a preset decision tree algorithm to train a sample user attribute vector, a sample user behavior index vector and a sample user; determining predicted behaviors of the target user according to the predicted behavior probability. The user behavior prediction method and device are applicable to prediction of user behaviors.

The invention provides a fast positioning method and system based on a multi-source sensor. The fast positioning method comprises the following steps: extraction and classification of motion features:utilizing EMD (Empirical Mode Decomposition) and FFT (Fast Fourier Transformation) to carry out energy spectrum analysis, carrying out feature classification by a classic decision-making tree model-NBC, so as to construct a pedestrian motion feature based on an accelerometer signal; recognition of a real-time proceeding state: utilizing feature classification of fuzzy least square support vectormachines to exactly estimate motion modes of pedestrians; parameter model estimation of gaits: calculating the step frequency and the step length of the pedestrians according to time difference of thepeak and the trough of the signal and the spectrum energy change, and by combination with the height and the weight of the pedestrians, initially establishing a step-length estimation model of the pedestrians; course estimation: realizing reliability estimation of pedestrian courses on the basis of a heuristic type drifting cancellation method. The fast positioning method and system provided by the invention have the beneficial effects that a pedestrian motion direction is accurately reckoned, high-accuracy indoor fast positioning is realized, and beneficial reference is provided for reliability application of an intelligent navigation system.

The invention relates to a fault data prediction method, a device, a computer device and a storage medium of a generator set. The method comprises: acquiring operating parameters in the current operation process of the generator set; carrying out decision classification on the operation parameters by using an established decision tree model; acquiring the output classification information after the decision tree model decision classification; outputting fault data prediction results based on the classification information. By using the fault data prediction method of the generator set, the accuracy of the fault analysis can be improved, and the occurrence of safety accidents can be reduced, therefore, the fault prediction accuracy is high.

The invention provides a user-side energy equipment intelligent maintenance method. The method comprises the following steps of obtaining the all-dimensional data of all energy equipment through an Internet of Things sensor system; adding the analog data into a digital model pre-stored in a cloud platform to obtain a digital virtual mirror image of the building energy system, and performing analogoperation and maintenance operation by operation and maintenance personnel based on the digital virtual mirror image to obtain the analog operation and maintenance data; extracting the key data fromthe simulated operation and maintenance data, carrying out abnormal point elimination on the key data, and analyzing the key data by utilizing a big data analysis algorithm to obtain relevant data ofthe equipment health degree portrait; carrying out the abnormal attribute classification according to the equipment health degree portrait to determine specific information of the fault hidden dangerand output corresponding alarm information, and using the cloud platform to update the fault decision tree model according to all data associated with the fault handling operation. The method has thebeneficial effects that the operation and maintenance cost is reduced, and the operation and maintenance efficiency is improved.