65results about How to "Accurately coded" patented technology

The invention discloses a recommendation method based on deep learning, belongs to the technical field of data mining, and solves the problems of an existing recommendation method that the potential factor vector of a project can not be predicted from text content information which contains the project descriptions and metadata so as to cause recommendation inaccuracy. The method comprises the following steps that: carrying out modeling on the implicit feedback characteristics of the historical behavior data of a user, and learning to obtain the implicit factor vectors of the user and the project after modeling; taking the implicit factor vector of the project as tag training to carry out modeling on the time sequence information of project text contents and deeply mine a network model; and for a new project which does not appear in the historical behavior data of the user, predicting the network model obtained through S(2) in the text content information of the project to obtain the implicit factor vector of the project, directly matching the implicit factor vector of the project with the implicit factor vector, which is obtained in S(1), of the user, and sorting matching degreesto obtain the new project recommendation list of each user. The method is used for recommending new projects.

The system and method of detection of low flying animals, such as birds, bats, and insects, and more particularly the detection of low flying animals using a radar system to detect the animals in three-dimensional airspace. The radar system produces narrowly focused radar pulses. The radar system comprises a single radar unit, an A/D proceeding apparatus, an A/D conversion apparatus, and a pan/tilt controlled base platform. The system and method further producing an avoidance response in an animal, and more particularly, producing an avoidance response by illuminating the animal with ultraviolet light.

The present invention provides a method (30) for training an artificial neural network (NN). The method (30) includes the steps of: initialising the NN by selecting an output of the NN to be trained and connecting an output neuron of the NN to input neuron(s) in an input layer of the NN for the selected output; preparing a data set to be learnt by the NN; and, applying the prepared data set to the NN to be learnt by applying an input vector of the prepared data set to the first hidden layer of the NN, or the output layer of the NN if the NN has no hidden layer(s), and determining whether at least one neuron for the selected output in each layer of the NN can learn to produce the associated output for the input vector. If none of the neurons in a layer of the NN can learn to produce the associated output for the input vector, then a new neuron is added to that layer to learn the associated output which could not be learnt by any other neuron in that layer. The new neuron has its output connected to all neurons in next layer that are relevant to the output being trained. If an output neuron cannot learn the input vector, then another neuron is added to the same layer as the current output neuron and all inputs are connected directly to it. This neuron learns the input the old output could not learn. An additional neuron is added to the next layer. The inputs to this neuron are the old output of the NN, and the newly added neuron to that layer.

The invention discloses an image retrieval method on the basis of multi-feature fusion, which is mainly used for solving the problem of low accuracy of retrieving an image in the prior art. The image retrieval method comprises the following implementing steps: (1) extracting three visual features and semantic attribute features of all images in an image set to be retrieved; (2) calculating an adjacent distance matrix among all the images in the image set to be retrieved; (3) carrying out rough retrieval on the inquired images in each feature channel; (4) according to a rough retrieval result of each feature channel, analyzing the semantic attribute features of the images in the rough retrieval result and selecting a reference image; (5) according to the selected reference image, calculating each feature fusion template matrix; (6) according to the obtained fusion template matrice, obtaining a fused distance measure matrix; (7) according to the obtained distance measure matrix, returning a retrieval result to a user. The image retrieval method obviously improves accuracy of final image retrieval and can be used for image retrieval.

A method, medium, and system encoding and/or decoding a multi-channel audio signal, and a method, medium, and system decoding a signal down-mixed from multi-channels to a 2-channel signal. The method of encoding an audio signal may include generating spatial cues indicating directivity information of a virtual sound source generated by at least two channel sound sources among a plurality of channels, and down-mixing the plurality of channel signals. The method of decoding an audio signal may include receiving inputs of spatial cues indicating directivity information of a virtual sound source generated by at least two channel sound sources among sound sources of a plurality of channels, and a signal down-mixed from the plurality of channel signals, and restoring the down-mixed signal to a plurality of channel signals by using the spatial cues. According to such systems, media, and methods, a multi-channel audio signal can be accurately encoded and/or decoded regardless of frequency bands.

The invention discloses a user classification method based on multi-model fusion, belongs to the technical field of machine learning, and solves the problems of less dimension, low precision and complex iteration in traditional model classification. The method comprises the steps that a data set containing user characteristics is acquired and processed; the method is based on an XGBoost algorithmand two different Y value intersection interval settings, different derivation methods are selected to derive characteristics in the processed data set; the data set features and all the derived features are input into two XGBoost models as a new data set for training and optimization, and the leaf node number of each training set sample at each tree is output by the two XGBoost models after optimization; one-hot coding is trained according to leaf node number ser, and the leaf node number output by each XGBoost model is subjected to the trained One-hot code conversion to obtain all LR features corresponding to the training set samples, and training an optimization LR model based on the LR features corresponding to all the training set samples; and an XGBoost + LR fusion model is obtainedto classify the users. The method is used for classifying users based on multi-model fusion.

A speech coding apparatus includes a spectrum parameter calculation section, an adaptive codebook section, a sound source quantization section, a discrimination section, and a multiplexer section. The spectrum parameter calculation section receives a speech signal and quantizes a spectrum parameter. The adaptive codebook section obtains a delay and a gain from a past quantized sound source signal using an adaptive codebook, and obtains a residue by predicting a speech signal. The sound source quantization section quantizes a sound source signal using the spectrum parameter. The discrimination section discriminates the mode. The sound source quantization section has a codebook for representing a sound source signal by a combination of non-zero pulses and collectively quantizing amplitudes or polarities of the pulses in a predetermined mode, and searches combinations of code vectors and shift amounts used to shift the positions of the pulses to output a combination of a code vector and shift amount which minimizes distortion relative to input speech. The multiplexer section outputs a combination of outputs from the spectrum parameter calculation section, the adaptive codebook section, and the sound source quantization section.

The embodiment of the invention discloses a decoding method which comprises the following steps: acquiring CNG parameters of a first frame of a first superframe from a speech coding frame at the frontof the first frame of the first superframe; carrying out background noise decoding to the first frame of the first superframe according to the CNG parameters, wherein the CNG parameters comprise target excitation gains and LPC filter coefficients, wherein the target excitation gains are determined by fixed codebook gains quantized by the long-term and smooth speech coding frame, and the LPC filter coefficients are defined by LPC filter coefficients quantized by the long-term and smooth speech coding frame. The invention also discloses a decoding device. By utilizing the embodiment, the occupied bandwidth can be remarkably reduced under the condition of ensuring the signal quality.

The invention discloses an electronic medical record and medical record home page diagnosis consistency verification method and system. The method comprises the steps: acquiring one or more diagnosisnames in medical record home page information; determining whether the diagnosis name meets a preset accepting and excluding condition of a pre-constructed international disease classification ICD10 coding characteristic knowledge base or not; grouping the diagnosis names according to the core diseases in response to the fact that the diagnosis names meet the accepting and excluding conditions; obtaining a first feature set of a core disease corresponding to each group of diagnosis from the ICD10 coding feature knowledge base; structuring the electronic medical record information according tothe structure of the ICD10 coding feature knowledge base; obtaining a second feature set of the core disease corresponding to each group of diagnosis from the structured electronic medical record information; sequentially comparing the acquired first feature set with the acquired second feature set according to the core disease; and outputting a verification result of whether the electronic medical record and the medical record home page diagnosis are consistent or not according to the comparison result. According to the invention, the consistency between the electronic medical record and themedical record home page diagnosis can be accurately verified.

The application discloses a medical record home page coding method and system. The method comprises the steps: responding to the existence of original names and original codes in inputted medical record home page information, and determining whether the original name matches an international disease classification ICD standard table or not for each original name; in response to the fact that the original name does not match the international disease classification ICD standard table, determining whether the original code meets a preset accepting and excluding condition of a pre-established ICDfeature knowledge base or not; in response to the fact that the original code meets a preset accepting and excluding condition, acquiring a code classification feature from an ICD feature knowledge base according to the original code and acquiring a name classification feature from the ICD feature knowledge base according to the original name; determining whether the name classification featuresmatch the code classification features or not; in response to matching of the name classification features and the code classification features, adding the original codes into a candidate code list; adjusting the coding content and/or sequence of the candidate coding list according to a pre-established coding rule knowledge base; and outputting the correct coding list. According to the invention,the coding correctness can be improved.

The invention relates to the technical field of big data, and discloses a word vector construction method and device, electronic equipment and a storage medium. A more accurate word vector is constructed by fusing a local word vector representing local context features and a topic distribution vector representing probability distribution of topics of texts where segmented words are located, so asto realize ambiguity elimination of polysemous words. The method comprises the steps: performing word segmentation processing on a plurality of to-be-processed texts to obtain segmented words in eachto-be-processed text; obtaining a local word vector of each segmented word based on the context of each segmented word; acquiring a topic distribution vector of each segmented word based on probability distribution of a topic to which each segmented word belongs in the plurality of to-be-processed texts; and respectively fusing the local word vector and the topic distribution vector of each segmented word to obtain a target word vector of each segmented word.

The invention discloses a coding method of a piece of ultra-thin glass. The coding method comprises the following steps: (1) virtual coding is carried out on pieces of ultra-thin glass arranged successively at a conveying mechanism and a monitoring point for monitoring passing by states of the ultra-thin glass is arranged at one side of the conveying mechanism; (2), passing time t0 of each piece of ultra-thin glass passing by the monitoring point and a time interval t between time of leaving of the previous ultra-thin glass and time of arriving at the monitoring point by the next ultra-thin glass are obtained; (3), a transmission speed of a conveying belt is read, and equivalent time t1 of passing by the monitoring point by each piece of ultra-thin glass is obtained based on the widths and spaces of the ultra-thin glass and the transmission speed; (4) according to the time interval and the equivalent time, whether a piece of ultra-thin glass misses is evaluated and the number of missing ultra-thin glass is also assessed; and (5), the virtual coding information and information of the number of the missing glass are stored and code pasting is carried out on the stored information by a code printing mechanism. Therefore, accurate coding of ultra-thin glass of about 1 mm can be realized; the quality can be improved; and the cost can be lowered.

The invention provides fluorescence coded microspheres based on long-life composite quantum dots and a preparation method thereof. The fluorescence coded microspheres based on long-life composite quantum dots comprise agarose microspheres, first quantum dots and second quantum dots; the first quantum dots and the second quantum dots are combined with the agarose microspheres by virtue of chemical bonds; each first quantum dot has a core-shell structure which takes cadmium telluride as a core and cadmium sulfide doped with copper ions as a shell; the doping molar fraction of the copper ions in the cadmium sulfide doped with the copper ions is 0.5%-2%, the size of the core is 1.8nm-2.2nm, the thickness of the shell is 2nm-4nm, the emission wavelength range of the first quantum dots is 700nm-910nm, and the fluorescent life is 0.8-1.2 microseconds; and the emission wavelength range of the second quantum dots is 700nm-820nm, and the fluorescent life is 100-220 nanoseconds. The invention further provides a preparation method of the fluorescence coded microspheres based on the long-life composite quantum dots.

The invention discloses microspheres for coding and decoding and a coding and decoding method and decoding system thereof. The coding and decoding method comprises the steps of S1, combining various nanometer materials containing metallic elements into microspheres to form various microspheres; S2, conducting laser inducing breakdown on one microsphere to be tested by means of pulse laser with pulse width at nanosecond scale and single pulse energy larger than 100 mJ; S3, receiving the plasma radiation spectrum generated by the microsphere to be tested after the microsphere is subjected to induced excitation; S4, conducting spectrum peak traversal in the plasma radiation spectrum, conducting comparison with standard spectrum peak data of a standard atomic spectrum, determining a data matched reference element as the metallic element in the nanometer material in the microsphere, and determining the microsphere to be tested as the microsphere containing the corresponding metallic element. Coding and decoding accuracy is high, and stability is high.

The invention discloses an automatic coding system and method for battery pack positions. The system is reasonably connected with one or more monitoring modules through a main control module via gradeconnecting lines or communication lines, and the one-time automatic coding function to a battery pack is achieved; and furthermore, due to reasonable wiring, the system has the advantages of being high in anti-interference capability and free of signal attenuation influences and is more stable and reliable in performance. According to the method, signal receiving and transmitting rules of the main control module and the monitoring modules are reasonably set to complete automatic coding of the battery pack; in the method, errors are not likely to occur in the signal transmitting process, and coding is accurate; and furthermore, the signal anti-interference capability is high, signal attenuation influences are avoided, and the method is fast and convenient to implement and reliable.

The invention discloses a message sending method, device and equipment and a message receiving method, device and equipment, and relates to the technical field of Internet of Vehicles, and the messagesending method comprises the steps: obtaining first data determined by a target coordinate; wherein the first data is an M-bit binary number; performing equal-length encryption on data of N bits in Mbits of the first data to obtain second data; wherein the N bits are preset low bits in the first data, M and N are both positive integers, and N is smaller than M; correspondingly updating the dataof the N bits of the first data into the second data to obtain third data of the target coordinates, the third data being data corresponding to the updated first data; and transmitting fourth data determined by the third data. According to the scheme, it can be ensured that the ciphertext after target coordinate encryption meets the input range of the encoder, accurate encoding is achieved, and the possibility of data transmission failure is reduced.

The embodiment of the invention belongs to the technical field of big data, is applied to the field of smart medical treatment, and relates to a medical data correction method and device, computer equipment and a storage medium. The method comprises the steps of receiving disease data which comprises disease names and disease codes; cleaning and merging the disease names to obtain a local disease name; respectively determining the number of each disease code associated with the local disease name, taking the disease code with the largest number as a local disease code, and generating a matching library based on the local disease code and the corresponding local disease name; and receiving a standard code library and medical data, wherein the medical data comprises to-be-coded names, the to-be-coded names are matched based on the standard code library and/or the matching library, target codes are obtained, and the target codes are associated with the to-be-coded names. The matching library may be stored in a blockchain. According to the invention, accurate coding of the to-be-coded name is realized so as to correct the original coded data of the to-be-coded name.