56 results about "Factor score" patented technology

Quantitatively improving a financial evaluation is disclosed. A plurality of factor scores based on a plurality of factors is determined. A total score is calculated based on the plurality of factor scores. A plurality of actions is identified that will influence at least a portion of the plurality of factor scores. The actions are ranked based on their total impact on the plurality of factor scores. A most important action of the plurality of actions to improve the total score is determined.

The invention provides a method for reducing dimensions of high spectroscopic data based on a factorial analysis model, comprising the following steps: (1) reading in the high spectroscopic data; (2) establishing the factorial analysis model for dimension reduction of the high spectroscopic data; (3) calculating average of data, covariance matrix and correlation matrix; (4) calculating proper value and standard eigen vector of the correlation matrix of the data; (5) carrying out solving on factor loading matrix by a parametric estimation method; (6) calculating covariance of special factors and communality of data variables in the factorial analysis model; (7) calculating the biggest factor loading spin matrix based on variance; (8) calculating factor scores by a least square method based on variance; and (9) obtaining eigen dimensionality characterizing high spectroscopic data, thereby realizing dimension reduction of the high spectroscopic data. The method is an automatic method for dimension reduction of the high spectroscopic data, which can effectively remove relativity of wave bands of the high spectroscopic data and improve separability of different categories of ground objects.

A software-product line analyzer including change history data, a change-history processing unit, a factor analysis unit, a variability analysis unit, and configuration information data, i.e., output data, wherein the change-history processing unit converts the change history data into numerical values, the factor analysis unit performing a factor analysis using inter-products change information as observation data, and, based on each factor and each factor score determined, the variability analysis unit determining variabilities in the software-product line and software components corresponding thereto, and outputting the configuration information data.

The invention relates to a mental health evaluation system based on an Internet. The system comprises the following steps: establishing and training a RBF neural network evaluation model based on the factor score of a mental test scale in a known sample stored in a cloud database by use of a RBF neural network algorithm; and acquiring the factor score of the mental test scale of a new individual, and obtaining a mental health state evaluation result of the new individual according to the RBF neural network evaluation model. The RBE neural network estimation model has good accuracy and adaptability to the evaluation of the mental health state, is small in error and good in fitting; the mental health result and therapeutic schedule obtained by use of the mental health evaluation system of the application are more precise and reliable.

The invention discloses a nonnegative-constrain-factor pollution source apportionment method based on naive Bayesian source identification. The nonnegative-constrain-factor pollution source apportionment method is used for apportionment of polycyclic aromatic hydrocarbon pollution sources in a water body, and includes the eight steps of determining a surveyed region; collecting base data; surveying the degrees of influences of pollution source industry enterprises on the environment; building fingerprint spectrums of polycyclic aromatic hydrocarbon pollutants in various emission sources; treating the fingerprint spectrums to obtain pollution source receptor samples; training the pollution source receptor samples with a naive Bayesian source method; carrying out calculation with a nonnegative-constrain-factor analysis method; finally, calculating the pollution source contribution rates of factor loads for main pollution source factors with high factor scores through a trained classification model, and achieving source apportionment of the polycyclic aromatic hydrocarbon pollutants. By means of the method, the sources of the polycyclic aromatic hydrocarbon pollutants in the water body can be rapidly and accurately traced, and powerful technical supports can be provided for making drainage-basin polycyclic aromatic hydrocarbon pollution control countermeasures and improving the quality of the drainage-basin water environment.

The invention discloses a user value classification method and device based on clustering analysis, and relates to the technical field of computers. The method specifically comprises the steps of determining an original feature vector of each user according to an index value under each index feature of the user; conducting factor analysis on the original feature vector of each user, and regarding public factors of which the factor scores satisfies a set rule as factor variables; determining a comprehensive feature vector of each user based on the value of each user under the factor variables; conducting clustering analysis on the comprehensive feature vector of each user to determine the classification of each user. According to the embodiment, index features can be prevented from being excessively complicated, the number of factors for clustering analysis is lowered, and the time cost and the database resource consumption amount for user value classification are reduced; the situation that the weight is given to each index feature according to experience can be avoided, the weight allocation accuracy is improved, and the accuracy of user value classification is then improved; the objectivity and the accuracy for user value grouping can be improved.

The invention provides an underground water type drinking water source pollution source analysis method, comprising preprocessing sample data; calculating feature values and feature vectors of the preprocessed sample data; solving a factor load matrix and a factor score matrix according to the feature value matrix and the feature vector matrix; carrying out nonnegative constraint rotation to the factor load matrix and the factor score matrix. The nonnegative constraint rotation is carried out to the factor loading matrix and the factor scoring matrix, negative values are prevented from appearing in the matrix decomposition result, the factor load and factor score are equipped with interpretability and clear physical significance; the pollution source and contribution are determined at the same time without knowing source profile in advance; the method is simple and convenient in operation; and measurement and detection uncertainty is effectively eliminated.

The present invention provides a quantitative trading strategy generation method based on multi-objective optimization, a device, a device, and a storage medium. The method includes: selecting target factors generated by the strategy; scoring each target factor on interval samples, and standardizing the scores , so that the score of the interval sample is in the defined interval; form the objective function, the objective function is composed of the objective factor and its corresponding weight, and the weight includes multiple groups; use the primary strategy to backtest the objective function, and generate the backtest result; select the expected Generate multiple groups of multi-objective weights based on the backtest results; select weights and adjust them according to investor preferences, and output optimal strategies. The present invention selects multi-target factors to form a quantitative strategy according to investor preferences, and further optimizes the strategy according to the preference weight among multiple groups of weights, so as to better reflect investors' investment preferences; evaluate the backtest results of the strategy by selecting a high score of a single target factor , reducing the number of strategies and the corresponding backtest running time, reducing the burden on system resources.

The invention discloses a high-dimensional evaluation method for oil field operation risk. Step 1) collect risk point data, and divide the risk point data according to four indicators: risk occurrence possibility, accident level, public opinion influence degree, and economic influence degree. 2) Select the risk points existing in the oilfield operation process to analyze the possibility and degree of influence, and draw the calculated data as a scatter diagram on the vertical axis of the degree of influence; 3) Use factor analysis to extract the original data Dimensionality reduction of effective information, calculation of factor scores of possibility and degree of influence, and drawing of factor risk matrix; 4) Quantitative determination of risk importance criteria by cluster analysis method, further optimization of risk matrix; 5) Reduction of existing data Dimensional quantitative analysis, using principal component analysis to complete risk ranking. The method described in the invention can be adapted to the operation risk assessment of different oil fields by adding pruning risk points, and the analysis results are accurate and suitable for popularization.

The invention discloses a mental health evaluation method based on Internet cloud database. The method comprises the following steps: establishing and training a RBF neural network evaluation model based on the factor score of a mental test scale in a known sample stored in a cloud database by use of a RBF neural network algorithm; and acquiring the factor score of the mental test scale of a new individual, and obtaining a mental health state evaluation result of the new individual according to the RBF neural network evaluation model. Through the adoption of the method disclosed by the invention, the RBF neural network and the Internet cloud database storage technology are combined to evaluate the mental health, the examination information is effectively used for improving the mental problem recognition rate and precisely matching the therapeutic schedule, and a more scientific, faster, more precise and comprehensive mental health state evaluation method is established.

A method for categorizing work including receiving a response to a questionnaire from a person. The questionnaire includes a plurality of scales, with each of the scales corresponding to one of a plurality of work characteristics. Each scale includes one or more questions related to the corresponding work characteristic. Scale scores are computed for each of the scales within the questionnaire using the response from the person and the questionnaire as input. Two or more scales are combined into a work factor. A work factor score is computed for the work factor based on the contribution of each of the combined scales to the work factor. A work category is assigned based on the factor score.

The invention discloses a talent evaluation method and a system based on an evaluation standard tree. The evaluation standard tree comprises a root node, a summary node and a content node, wherein thesummary node is a non-leaf node, comprising summary rules; the content node is a leaf node, comprising a summary rule and an element score, and the element score records the corresponding relationship between the condition and the score; The evaluation method comprises the following steps: receiving the configuration of each node in the evaluation standard tree by the user and the information list of the person to be evaluated; the score of the person to be evaluated being calculated according to the evaluation standard tree; the score of each content node being calculated according to the information of the person to be evaluated and the factor score of each content node; according to the summary rules of each content node and summary node, summarizing level by level to obtain the totalscore of the person to be evaluated. Through the evaluation standard tree, the invention can flexibly configure the evaluation standard and quickly calculate the score of the person to be evaluated.

An incoming order is matched or allocated to trade with a plurality of resting orders. Order book data indicative of the resting orders is obtained. For each resting order, a set of order quality factor scores is determined based on the order book data. The order quality factor scores include any combination of two or more of a first factor score indicative of order quantity, a second factor score indicative of order book position, and a third factor score indicative of order duration without modification. A ranking of the plurality of resting orders is determined based on the set of order quality factor scores determined for each order of the plurality of resting orders. A volume of the incoming order is allocated across a subset of orders of the plurality of resting orders based on the ranking in partial satisfaction of the incoming order.

The invention belongs to the technical field of tobacco processing and detection, and particularly relates to a method for detecting a blending proportion of components in cut tobacco. The method comprises steps of pretreating samples of the components of the cut tobacco, building a calibration set sample, constructing a blending proportion prediction model based on synthetic chemical factor scores, establishing a near-infrared prediction model of synthetic factor scores, performing detection and judgment and the like. According to the method for detecting the blending proportion of the components in the cut tobacco, multiple synthetic chemical factors are taken as chemical characteristic values for characterizing the blending proportion of the cut tobacco, and evaluation requirements forblending uniformity of different types and different categories of the cigarette cut tobacco can be met better. With adoption of the method for detecting the blending proportion of the components in the cut tobacco, not only is the foundation laid for detection and judgment of the components of the cut tobacco, but also the good foundation can be laid for online continuous monitoring and control for blending uniformity of the cut tobacco, and the method has better practical value and popularization and application significance.

The invention discloses a customer classification method and system, and relates to the field of customer classification, and the method comprises the steps: obtaining historical data; generating a corresponding feature dimension data table according to the historical data and a preset feature dimension index, the feature dimension data table comprising feature dimension data of a plurality of customers; establishing an analysis sample pool according to the feature dimension data table; carrying out dimension reduction on the feature dimension data in the analysis sample pool by adopting a factor analysis algorithm to obtain a corresponding factor score data table; and according to the factor score data table, classifying the customers by adopting a clustering algorithm. The dimensionalityreduction effect of the factor analysis algorithm and the clustering algorithm are adopted for classification at the same time, the limitation that the number of feature dimensionalities can be considered in a conventional customer classification method is broken through, and the two algorithms are combined for use, so that the precision and the accuracy of the classification result are improved.

The invention provides a method for developing typical working conditions of an energy storage battery. The method comprises the following steps: determining an actual operation working condition performance parameter matrix of the energy storage battery; determining a factor analysis mathematic model and solving a factor loading matrix; computing the distance of factor scores of any two working condition fragments; reconstructing the typical working conditions of the energy storage battery. The method for developing typical working conditions of the energy storage battery can be used for comprehensively, correctly and pointedly grasping the working condition applicability of the energy storage battery, and is simple, feasible and easy to popularize.

A method for characterizing the difference degree of cigarette style characteristics by using indicators comprises the following steps: screening the chemical detection indicators of finished cigarette cut tobacco, and determining twenty representative indicators as characteristic chemical indicators; selecting cigarette samples-to-be-characterized having different style characteristics, and detecting the twenty characteristic chemical indicators of the selected samples; performing R-factor analysis on detection indicator data of all the samples, and calculating all factor scores and the factor comprehensive scores of every sample and the factor comprehensive scores of every style characteristic cigarette; and quantitatively characterizing the difference degree of two cigarette style characteristics by using the absolute value of a difference between the factor comprehensive scores of the two style characteristic cigarettes, DI. The evaluation method can realize the quantitative characterization of the difference degree of the cigarette style characteristics, and can provide guidance for cigarette quality maintenance and quality control.

The invention belongs to the technical field of wine making, and in particular relates to a method for quantitatively evaluating quality of bases of strong flavor Chinese spirits. According to the method for quantitatively evaluating quality of the bases of the strong flavor Chinese spirits, results are reasonable and accurate. The method includes screening n common characteristic peaks with a peak area X in m write spirits; forming areas the common characteristic peaks in m write spirits into a matrix R, inputting the R into a statistic package for social science (SPSS) to perform principal component analysis, extracting s principal components, a variance contribution rate of the s principal components, and a factor score coefficient matrix D according to analysis results; obtaining a standardized linear combination Fi=d1iX1+d2iX2+d3iX3+......dniXn (i=1, 2, ......, s) of the s principal components according to the matrix D, performing regression analysis by taking G as independent variable g for sum of products of each principal component and the variance contribution rate h, and the white spirits quality grade score U as a dependent variable to obtain a plurality of regression equations U=f (g), and taking the highest regression equation of coefficient of determination R2 as a final mathematical model evaluation equation.

The invention discloses a medium and long term power load forecasting method based on a factor-main attribute model. The medium and long term power load forecasting method based on the factor-main attribute model includes that step 1, building (n-1) impact factor indexes X1-Xn-1 and an original matrix of a forecast object Xn, using Z standardization to pre-process the impact factor index data X1-Xn-1 and Xn, and carrying out dimensionless on the index data to obtain an index matrix Ao*n; step 2, according to a factor analyzing method, confirming a selected common factor, calculating the corresponding factor score, and building a factor forecasting model FORMULA (shown in the description); step 3, sieving to obtain main attributes m1, ..., mr according to a main attribute algorithm, wherein the main attribute model is FORMULA (shown in the description), and v is the feature vector; step 4, building a factor-main attribute medium and long term power load forecasting model according to FORMULA (shown in the description), calculating to obtain a normalized forecast variable value, performing Z standardization formula conversion, and calculating again to obtain the practical value of the corresponding original variable value.

The invention provides a system and method of a selecting an egress connection from egress connections at a node for a call being established in a network along a path being associated with the node. Each connection has sufficient bandwidth to carry the call and terminates at a same destination point in said path. The system and method comprise utilizing at least a bandwidth load balance value for each connection in a set of the connections to select the egress connection from the set. An administrative factor score may be used to further define members of the set.

The invention relates to a human body physique clustering method based on a factor analysis and a probabilistic neural network. The method comprises the following steps of collecting static human bodynet body data and constructing a human body database; determining 27 basic variables according to the actual and research needs of the clothing industry; using the factor analysis to reduce the dimensionality of the 27 basic variables, and according to the cumulative contribution rate of principal components, extracting the factor scores of the top N principal components with contribution rate greater than 80%; constructing a probabilistic neural network model of human body physique clustering, wherein the factor scores of the top N principal components is an input layer, and an output layeris the category code of a human body physique; clustering the human body physique based on the factor scores and the probabilistic neural network. The method can directly and effectively classify human body physiques.

The invention discloses a factor score-based tobacco sales quantity prediction method. The method comprises the steps of collecting data of a plurality of independent variable factor groups to establish a database; obtaining corresponding factor scores through methods of data quantitative processing, centralization processing, Gram-Schmidt transform, collaborative filtering and the like; and according to weights of factors, predicting a sales quantity. According to the method, by utilizing an internet platform, relations with tobacco retailers and consumers can be established, and consumer information can be obtained in time; various influence factors are subjected to score value assignment, so that the sales quantity of cigarettes is predicted, accurate marketing is achieved, and tobacco industry reform is facilitated.