38results about How to "Improve scoring accuracy" patented technology

The invention discloses a recommendation method based on singular value decomposition and classifier combination. The recommendation method comprises the steps of computing average score and probability distribution of an item through data preprocessing; training a singular value decomposition model through a stochastic gradient descent method, computing an entropy set of a scored item set of the user in item classification through a computing method of entropy, and determining an uncertainty critical value of the item; and comparing and predicting uncertainty and critical value of the item to determine whether to use a classifier, and recommending N items with highest scores in all non-scored items of the user through a Top-N method. According to the method, individual recommendation is produced on the basis of analysis of historical score data of the user; predicting score of a designated item i is acquired through a singular value decomposition algorithm, information entropy of the item for each user is calculated so as to determine whether to classify, and final prediction score of the item is acquired through the classifier, so that the accuracy of the recommendation method is improved.

The invention discloses a subjective question scoring model and scoring method based on deep learning BERT-CNN. The subjective question scoring model comprises a BERT conversion word vector matrix module, a CNN semantic feature extraction module, a similarity calculation module and a scoring module, wherein the BERT conversion word vector matrix module is used for converting an answer text into aword vector matrix, and then transmitting the word vector matrix of each answer to the CNN semantic feature extraction module; the CNN semantic feature extraction module is used for acquiring a semantic feature vector of the answer text, includes a convolution layer and a pooling layer, and is used for obtaining semantic feature vectors of the answer texts and then transmitting the signals to thesimilarity calculation module for calculating a similarity value between the answer text semantic feature vectors and transmitting the similarity value to the scoring module for determine the score ofthe answer text. The subjective question scoring model is applied to subjective question scoring, can effectively reduce the labor cost, and can solve the problems that due to the fact that only keyword matching is carried out in manual scoring, the scoring result is inaccurate, and scoring is not fair.

Methods, systems and kits to detect and/or quantify disease progression in a subject having a progressive form of MS are disclosed.

This invention mainly discloses an EOG-based sleep staging method which is used to improve of an accuracy of automatic sleep judging result. The method may be executed by a processor and comprises steps of reading a to-be-test data having a variation depending on time of EOG generated by a user during a sleep process; acquiring a plurality of eye movement characteristics based on a variation of the EOG of a period, with the characteristics including an eye movement ratio, a blink count, a low frequency power ratio, a high frequency power ratio, an alpha rhythm ratio, a spindle rhythm ratio, a delta rhythm ratio and an average amplitude of eletrooculogram (EOG) signals; and determining a sleep state of the period is a “WAKE stage”, a “REM stage”, a “S1 stage”, a “S2 stage”, or a “SWS stage” based on the EOG characteristics. Furthermore, a computer readable medium with stored programs, and electronic apparatuses are provided.

The embodiments of the invention relate to a logistics scheduling method and a logistics scheduling device. The method comprises the following steps: acquiring a first rule scoring model in a logistics scheduling system; calling a preset machine learning model to train and fit the first rule scoring model to get a first machine learning model; training the first machine learning model based on preset historical order data to get a second machine learning model; and taking the second machine learning model as a rule scoring model in the logistics scheduling system. Iteration continues until a second machine learning model tending to be stable is obtained. Moreover, the second machine learning model is called to score all relevant links in the logistic scheduling system. The accuracy of scoring can be improved greatly, and the efficiency of logistics distribution can be improved.

The invention discloses a game assessment method, device and equipment. The method comprises the steps that the score of each dimension of the assessed game of users participating in assessment is acquired; and average calculation is performed according to the sum of the scores of all dimensions and the number of total dimensions so as to determine the user dimension score value. According to thescheme, multidimensional scoring of the game can be realized so that the more comprehensive and more accurate reference information can be provided for game selection for the users.

The embodiment of the invention provides a chromosome metaphase image scoring method and device, electronic equipment and a storage medium, and the method comprises the steps: obtaining N chromosome metaphase images, N being an integer greater than 1; respectively inputting each chromosome metaphase image into a trained neural network to obtain a plurality of image features corresponding to each chromosome metaphase image; and according to the plurality of image features corresponding to each chromosome mid-term image, each chromosome mid-term image is scored. According to the embodiment of the invention, each feature in each chromosome metaphase image is extracted by using the trained neural network, so that the objectivity of a scoring basis is ensured, the scoring accuracy is improved, scoring can be carried out from the perspective of multiple features, metaphase image selection habits of different users can be met, and the user experience is improved. And a user can conveniently select medium-term images meeting requirements from the images.

The invention discloses a user credit scoring method, a storage medium and a server. The user credit scoring method comprises the steps of: providing user basic data; generating fact labels accordingto data characteristics of the user basic data; generating at least one model label according to the fact labels; and scoring each preset prediction label by adopting a weighting model according to the plurality of fact labels and the at least one model label. According to the user credit scoring method, the user basic data is acquired through multiple channels, the basic data acquired through thechannels is integrated and analyzed to obtain the fact labels and the model labels, finally, the preset prediction labels are scored in a weighted manner, and the user credit score is determined according to the score of each prediction label, so that the scoring accuracy is improved.

The invention discloses an air purifier demonstration and scoring method and device. The method comprises the following steps of gas distribution, testing, subjective scoring and data comprehensive processing, and finally a comprehensive purifying performance index P value of an air purifier is obtained, wherein P is no less than zero and no more than 1, the smaller the P value is, the poorer the purifying performance is, and the larger the P value is, the higher the purifying performance is. The calculation method of the P value is shown in the specification. The air purifier demonstration and scoring method has the advantages that the demonstration result is relatively comprehensive, and the scoring accuracy is relatively high. The invention further discloses an air purifier demonstration and scoring device.

The invention provides an enterprise scoring model construction method, an enterprise scoring method and an enterprise scoring device, and the enterprise scoring model construction method comprises the following steps: obtaining portrait data and operation data of the same kind of enterprises, carrying out dimension reduction processing on the portrait data and the operation data, and selecting anin-mold index from a dimension reduction result; Classifying the in-mold indexes, and constructing class indexes; For the class indexes, executing the following steps: determining a judgment matrix according to an important relationship among the in-mold indexes contained in the class indexes; Performing normalization processing on the judgment matrix, and determining a judgment matrix weight data set corresponding to the judgment matrix; Calculating a variance weight data set corresponding to the class index by using a variance method; And calculating an in-mold index weight according to thejudgment matrix weight data set and the variance weight data set. According to the scheme provided by the invention, the influence of human interference or subjective factors is greatly reduced, so that the scoring accuracy is effectively improved.

The invention discloses ocean engineering rock mass quality scoring methods based on simple testing and fine testing. Aiming at the particularity of ocean rock mass engineering, the invention providestwo specific ocean engineering rock mass scoring methods, and the method comprises the steps: firstly, carrying out on-site point load testing, direct observation and longitudinal wave velocity testing to quickly obtain a preliminary ocean engineering rock mass quality score; subsequently, performing a uniaxial compressive strength test, a quantitative statistical test, a qualitative descriptiontest, a longitudinal wave velocity test, a rock disintegration resistance test and other fine tests on ocean engineering rock mass to form a more accurate oceanographic engineering rock mass quality score. The method has the advantages that the method has sufficient theoretical basis and practical value, rapid, simple and convenient testing and accurate fine testing are combined, reliable scoringparameters and technical support are provided for foundation design and construction of ocean rock mass engineering, operability is high, and the obtained scores are real and reliable.

The invention relates to a dynamic extensible voice quality inspection scoring method. The method comprises the steps of acquiring recording audio data of an object subjected to quality inspection, converting the acquired recording audio data into text data in a text format, converting the text data into a voice oscillogram, detecting service taboo languages according to the text data, generatinga first score deduction value; and detecting emotion abnormality according to the voice oscillogram, generating a second score deduction value, obtaining a quality inspection score of the recorded audio data according to the first score deduction value, the second score deduction value and the full score, and feeding back a judgment result of the quality inspection score, service taboo languages existing in the text data and whether the recorded audio data has abnormal emotion or not. The voice quality inspection scoring method is an automatic quality inspection scoring method, does not need to specially set quality inspection personnel, further avoids the problem of large quality inspection workload, reduces personnel investment cost, avoids loss of scoring fairness due to subjective judgment of the quality inspection personnel, and is relatively high in accuracy.

The invention relates to an auxiliary judgment system using angle big data matching, and the system comprises a database application node which is used for storing each degree of completion of each angle corresponding to the set number of jumping turns of each athlete through a database; the angle identification component is used for calculating the surrounding angle of each single-point curve and removing the maximum value and the minimum value of each surrounding angle corresponding to each received single-point curve to obtain a plurality of residual angle values; and the action recognition part is used for analyzing the residual angle value with the highest occurrence frequency in the plurality of residual angle values as an effective angle value, and judging the completion degree of the set jump turns of the current athlete in the skating competition based on the effective angle value. According to the invention, the database can be adopted to store the corresponding relationship between the angle and the completion degree, and the pixel point-level single-point curve probability is introduced to help a referee provide an electronic auxiliary judgment mechanism for the completion degree of the skiing cycle number of a skate, so that the probability of misjudgment of the referee is reduced.

The invention provides a picture screening method, system and equipment, and a storage medium. The method comprises the following steps: obtaining to-be-tested pictures; scoring the to-be-tested pictures by using the trained target deep learning model to obtain a score value of each to-be-tested picture, wherein the target deep learning model is obtained by training an initial deep learning model based on a second training set, the initial deep learning model is obtained by training a preset deep learning model based on a first training set, the second learning rate in the initial deep learning model training process is smaller than the first learning rate in the preset deep learning model training process, and the preset deep learning model is a Transform model; and screening the to-be-tested pictures according to the score values. The accuracy of screening the tourism display pictures is improved.

The invention relates to an enterprise risk scoring method and device, computer equipment and a storage medium. The method comprises the steps of obtaining enterprise index data corresponding to a to-be-scored enterprise; inputting the enterprise index data into a constructed enterprise risk scoring model; obtaining an enterprise risk score corresponding to the to-be-scored enterprise output by the enterprise risk scoring model, wherein the enterprise risk score is used for measuring the risk condition of the scored enterprise; wherein the enterprise risk scoring model obtains a correspondingenterprise index importance score according to the enterprise index data; and according to the enterprise index importance score, constructing and obtaining an enterprise influence factor, according to the enterprise influence factor, calculating and obtaining an enterprise index weight, and according to the enterprise index data and the enterprise index weight, obtaining an enterprise risk scorecorresponding to the to-be-scored enterprise. The scoring accuracy of the enterprise risk scoring model can be improved.

The invention relates to the technical field of automatic driving, and discloses a lane line correlation degree determination method and device, equipment and a storage medium, which are used for improving the scoring accuracy of a correlation lane line between every two frames of lane line detection image data when a deviation distance exists between lane lines of different frames and the lane lines are relatively staggered. The lane line correlation degree determination method comprises the following steps: acquiring multiple frames of lane line detection image data, wherein each frame of lane line detection image data comprises multiple detected lane lines; performing lane line discrete point association processing on the multiple frames of lane line detection image data to obtain a plurality of lane line association degree scores between every two frames of lane line detection image data; and determining an associated lane line between every two frames of lane line detection image data according to the plurality of lane line association degree scores.

The invention relates to a method and device for detecting the yield of injection molding products and electronic equipment. The method comprises the steps: obtaining production parameters of an injection molding machine and temperature data and pressure data in an injection mold; preprocesing the production parameters of the injection molding machine and the temperature data and the pressure data in the injection mold; and inputting the preprocessed production parameters of the injection molding machine, the temperature data in the injection mold and the pressure data into the trained deep learning network, and outputting the yield score of the injection molding product through the deep learning network. According to the scheme, the injection molding product yield is scored through the deep learning network, so that the injection molding product can be detected in real time, and bad injection molding products are prevented from entering a subsequent process flow.