168 results about "Decision boundary" patented technology

The invention is a computer implemented technique for id entifying anomalous data in a data set. The method uses cascaded k-Means clustering and the ID3 decision tree learning methods to characterize a training data set having data points with known characterization. The k-Means clustering method first partitions the training instances into k clusters using Euclidean distance similarity. On each training cluster, representing a density region of normal or anomaly instances, the invention builds an ID3 decision tree. The decision tree on each cluster refines the decision boundaries by learning the sub-groups within the cluster. A test data point is then subjected to the clustering and decision trees constructed form the training instances. To obtain a final decision on classification, the decisions of the k-Means and ID3 methods are combined using rules: (1) the Nearest-neighbor rule, and (2) the Nearest-consensus rule.

A method comprising computer receiving at least one of time and location-referenced state data for an object of interest, determining present location of a vehicle within two presently overlapping fat paths, fat paths comprising homotopically distinct regions of travel, determining distance of vehicle from a point of divergence of fat paths, fat paths diverging to avoid object, the computer generating a decision boundary reachable prior in time to point of divergence wherein decision boundary is in advance of the present location of vehicle, computer generating a first second set of feasible headings for the vehicle, the first and second set respectively associated with a projected first and second crossing points of the decision boundary by vehicle wherein feasible headings promote positioning of vehicle in one of fat paths beyond point of divergence, and computer sending first and second sets of feasible headings to vehicle prior to vehicle reaching decision boundary.

The invention discloses a class-imbalance problem classification method based on expansion training data set; the method comprises the following steps: obtaining a true data set needed by a classification task; screening a few class samples from the true data set, and distinguishing samples that are close to and far away from the decision boundary; inputting said samples, running a productive confrontation network, thus obtaining artificial samples similar to the true data; adding certain amount of artificial samples into the true data set, thus obtaining a mixed data set; inputting the mixeddata set, and using a classifier to classify. The method combines a CycleGAN model with the boundary information of an original data set, thus effectively simulating distribution features of the truedata. The method samples small sample data so as to improve the classifier precision, and effectively preventing the class-imbalance problem from affecting the classification task.

A method for training a classifier for use in a computer aided detection system includes providing a training set of images acquired from a plurality of patients, each said image including one or more candidate regions that have been identified as suspicious by a candidate generation step of a computer aided detection system, and wherein each said image has been manually annotated to identify lesions, using said training set to train a classifier adapted for identifying a candidate region as a lesion or non-lesion, clustering candidate regions having similar features for each patient individually, and modifying said trained classifier decision boundary with an additional classification step incorporating said individual candidate region clustering.

A method is provided for determining a state-of-health of a battery in a vehicle-during an engine cranking phase. An engine cranking phase is initiated. Characteristic data is recorded that includes battery voltage data and engine cranking speed data during the engine cranking phase. The characteristic data is provided to a pre-processing unit. The pre-processing unit normalizes the characteristic data for processing within a classifier. The normalized data is input o the classifier for determining the vehicle battery state-of-health. The classifier has a trained state-of-health decision boundary resulting from a plurality of trials in which predetermined characterization data is collected with known classes. The battery state-of-health is classified based on the trained state-of-health decision boundary.

A computer-implemented method, apparatus and computer program product for projecting a machine learning model, the method comprising: obtaining a computerized multi-dimensional unsupervised anomaly detection model; obtaining a probability density function of the anomaly detection model; determining samples of the anomaly detection model, based on the probability density function; projecting the samples over at least one dimension set to obtain projected samples; processing the projected samples to obtain decision boundaries of the anomaly detection model over the at least one dimension set; and providing a visual display of the decision boundaries on a display device.

Systems and methods are described in relation to specific technical improvements adapted for machine learning architectures that conduct classification on numerical and/or unstructured data. In an embodiment, two neural networks are utilized in concert to generate output data sets representative of predicted future states of an entity. A second learning architecture is trained to cluster prior entities based on characteristics converted into the form of features and event occurrence such that a boundary function can be established between the clusters to form a decision boundary between decision regions. These outputs are mapped to a space defined by the boundary function, such that the mapping can be used to determine whether a future state event is likely to occur at a particular time in the future.

The welding quality classification apparatus relating to the present invention is an apparatus, wherein a data point indicating feature information of a welded joint to be classified whose welding quality is unknown is mapped to a point in a mapping space which has a dimensional number higher than the number of the features constituting the feature information, and the welding quality of a welded joint to be classified is classified based on which of regions of two welding qualities, which are formed by separating the mapping space with a decision boundary, contains the mapped point, and wherein a discriminant function is determined by adopting a weight which minimizes the sum of the classification error corresponding to classification accuracy of a training dataset and a regularization term having a positive correlation with the dimensional number of the discriminant function as weight for each feature constituting the discriminant function indicating the decision boundary.

The invention relates to the technical field of data mining, and discloses a selective up-sampling combined method for weighted ensemble classification prediction of unbalanced data flows. The method comprises the following steps of: screening minority class samples of history data blocks according to a similarity, and selecting samples closest to the current training data block in the aspect of concept; synthesizing the selected samples into new samples in a decision boundary area so as to selectively implement up-sampling; and carrying out weighted ensemble classification on the new sample by adoption of a probability distribution relevancy-based weight distribution strategy. According to the method, the minority class sample information is effectively increased through selecting history data with high similarities and synthesizing new data at the boundary area, so that the decision domain of the minority class is enlarged; and meanwhile, in order to adapt the dynamic data with concept drift and use an ensemble classification thought, the probability distribution relevancy-based weight distribution strategy is designed, so that the overall classification precision is enhanced. Experiment results show that the method is capable of effectively improving the minority class identification rate and the overall classification performance, and has the advantage of better processing the unbalanced data flows.

A system for generating an adversarial example in respect of a neural network, the adversarial example generated to improve a margin defined as a distance from a data example to a neural network decision boundary. The system includes a data receiver configured to receive one or more data sets including at least one data set representing a benign training example (x); an adversarial generator engine configured to: generate, using the neural network, a first adversarial example (Adv1) having a perturbation length epsilon1 against x; conduct a search in a direction (Adv1-x) using the neural network; and to generate, using the neural network, a second adversarial example (Adv2) having a perturbation length epsilon2 based at least on an output of a search in the direction (Adv1-x).

The invention relates to an apparatus and method avoiding collision with obstacles in automatic parking assistance system. According to the invention, a plurality of sensors are arranged in the front part, the rear part, and the side parts of a vehicle, so that the moving detection area of the obstacles is enlarged, an obstacle area in the prior art which detects the obstacles in all of the images and only detects the images having brightness changes is replaced, the movement of the obstacles is detected by means of a main component analyzing method and an obstacle image characteristic valve, and a support vector machine is employed to indicate the decision boundaries by means of some of support vectors. According to the support vector machine, the detection performance is improved because the movement of the obstacles is detected by the main component analyzing method and the obstacle image characteristic value; the obstacle detection and the movement can be automatically detected without the need for drivers' operation, thereby being convenient; and the minimal change of a device in prior art can be realized by adding the functions of obstacle movement detection and collision avoidance.

The invention provides a handwritten form identification system of BP neural network based on a dynamic sample selection strategy. Weights of different layers of network neurons are initialized randomly; a gradient descent method is used to optimize the network weight, in first round of iteration, all samples are used to calculate the total gradient, the total gradient is used to update the weights of different layers, and whether a sample serves as a training sample in next round of iteration is determined according to whether the sample is far from a decision boundary, and the training samples selected in the last round are used to calculate the total gradient, update the weights of different layers and select samples for next round of iteration repeatedly till the minimal stop error or the maximal interaction frequency is reached; and the obtained neural network is used to identify an unknown hand-written font sample. Compared with a traditional classification technology, According to sample selection strategy of the invention, the samples are selected dynamically according to the distances to the decision boundary, the amount of training sample is decreased step by step, and an algorithm can effectively solve the problem that training time of the BP network is too long in a big data set.

A dataset including at least one temporal event sequence is collected. A one-class sequence classifier f(x) that obtains a decision boundary is statistically learned. At least one new temporal event sequence is evaluated, wherein the at least one new temporal event sequence is outside of the dataset. It is determined whether the at least one new temporal event sequence is one of a normal sequence or an abnormal sequence based on the evaluation. Numerous additional aspects are disclosed.

Method, system, and program product for measuring audio video synchronization. This is done by first acquiring audio video information into an audio video synchronization system. The step of data acquisition is followed by analyzing the audio information, and analyzing the video information. Next, the audio information is analyzed to locate the presence of sounds therein related to a speaker's personal voice characteristics. The audio information is then filtered by removing data related to a speakers personal voice characteristics to produce a filtered audio information. In this phase filtered audio information and video information is analyzed, decision boundaries for Audio and Video MuEv-s are determined, and related Audio and Video MuEv-s are correlated. In Analysis Phase Audio and Video MuEv-s are calculated from the audio and video information, and the audio and video information is classified into vowel sounds including AA, EE, OO, silence, and unclassified phonemes. This information is used to determine and associate a dominant audio class in a video frame. Matching locations are determined, and the offset of video and audio is determined.

An improved decoding technique useful for hard decision decoding, such as quadrature phase shift keying (PSK) and quadrature amplitude modulation (QAM), as well as soft-decision techniques, such as Viterbi decoding and trellis decoding. The system in accordance with the present invention provides adaptive decision regions for hard-decision decoding techniques and adaptive metrics for soft-decision detection techniques in which the decision boundaries and reference constellations, respectively are optimized in order to minimize the bit error rate (BER). In particular, the decision boundaries and metrics are optimized based on the locations of the received constellation points. By adaptively adjusting the decision boundaries and metrics, the BER can be greatly improved with the need for nonlinear predistortion at the transmitter thus reducing the hardware complexity and weight of the transmitter which provides additional benefits in applications, such as satellite communication systems, where the transmitter is located on the satellite.

This invention relates generally to a system and method for classifying input patterns into two classes, a class-of-interest or a class-other, utilizing an Adaptive Fisher's Linear Discriminant method capable of estimating an optimal Fisher's linear decision boundary for discriminating between the two classes, when training samples are provided a priori only for the class-of-interest. The system and method eliminates the requirement for any a priori knowledge of the other classes in the data set to be classified. The system and method is capable of extracting statistical information corresponding to the “other classes” from the data set to be classified, without recourse to the a priori knowledge normally provided by training samples from the other classes. The system and method can re-optimize (adapt) the decision boundary to provide optimal Fisher's linear discrimination between the two classes in a new data set, using only unlabeled samples from the new data set.

Provided a secure pattern recognition method. The method includes: receiving data and generating a probe by converting the received data into a template for pattern recognition; accessing a gallery that is a template registered and stored in advance; determining a region to which the probe belongs and obtaining the center point of the region; obtaining a hash value of the center point and coordinate of the probe; and determining whether or not the hash value of the center point and a hash value stored in the gallery are equal and determining whether or not the probe and the gallery are classified into the same class by calculating whether or not the coordinate of the probe is inside a decision boundary configured with thresholds on the basis of the coordinates of the center point.

Method, system, and program product for measuring audio video synchronization. This is done by first acquiring audio video information into an audio video synchronization system. The step of data acquisition is followed by analyzing the audio information, and analyzing the video information. In this phase audio and video information is analyzed, decision boundaries for Audio and Video MuEv-s are determined, and related Audio and Video MuEv-s are correlated. In Analysis Phase Audio and Video MuEv-s are calculated from the audio and video information, and the audio and video information is classified into vowel sounds including AA, EE, OO, silence, and unclassified phonemes This information is used to determine and associate a dominant audio class in a video frame. Matching locations are determined, and the offset of video and audio is determined.

Method and apparatus for receiving coded signals with the aid of CSI are provided. The method comprises: performing channel estimation to obtain a CFR estimation vector; computing a squared magnitude of the CFR estimation vector, and obtaining a normalization factor α by averaging the squared magnitudes of CFR estimations on all N subcarriers; finding a norm-shift operand m satisfying the condition that α₀=2^m is a power of 2 number closest to the normalization factor α; performing a CSI-aided one-tap channel equalization on an output signal vector from a DFT processor by using the norm-shift operand m; performing constellation demapping; and performing channel decoding. The method further comprises obtaining a weighting factor vector by right shifting m bits of the squared magnitude of the CFR estimation vector so that the constellation demapping can use the weighted decision boundary values in case that its input signal is sensitive to both amplitude and phase.

Being applicable to solve the issue that in current technique, demodulation procedure is too complex, the invention discloses simplified schemes for two algorithms. Scheme 1 is that using LLR algorithm and MAX-log algorithm simplifies current demodulation procedure into piecewise functions. Scheme 2 is that using minimum distance between the received signal and hard decision boundary determines amplitude of soft bits to simplify demodulation procedure further. First, the invention decomposes signal in orthogonal into first branch component, and second branch component, and then calculates their soft bits so as to accomplish soft decision-making procedure. In scheme 1, the equipment includes communication interface (CI), decomposition module (DM), judgment module (JM), computing module (CM), and output module (OM). In scheme 2, the equipment includes CI, DM, JM, CM, and OM.