46 results about "Predictive variables" patented technology

A computerized method includes analyzing data associated with a credit line during an origination stage for predictive variables for use in a model for first party fraud, and flagging an account during the origination stage when at least one or more predictive origination stage variables cause a model score to exceed a pre-defined fraud likelihood threshold. The computerized method also includes analyzing data associated with one or more previously flagged, post-booked stage credit lines for data elements or transactions to be used as variables in a model to predictive of first party fraud at the customer-level or in one or more of the post-booked stage credit lines.

Parallelize a computer program by scoping program variables at compile time and inserting code into the program. Identify as value predictable variables, variables that are: defined only once in a loop of the program; not defined in any inner loop of the loop; and used in the loop. Optionally also: identify a code block in the program that contains a variable assignment, and then traverse a path backwards from the block through a control flow graph of the program. Name in a set all blocks along the path until a loop header block. For each block in the set, determine program blocks that logically succeed the block and are not in the first set. Identify all paths between the block and the determined blocks as failure paths, and insert code into the failure paths. When executed at run time of the program, the inserted code fails the corresponding path.

A system and method for pricing of derivatives is presented. A volatility clustering time series process, including one or more predictive variables, is generated with an innovation process. Marginals of a probability distribution for the time series process follow a smoothly truncated heavy tailed and asymmetric probability distribution. Model parameters for the time series process are calibrated to a set of exogenously provided derivative prices. Pricing of derivatives, including options and swaps, is determined.

Methods and systems for determining the importance of each of the variables, or combinations of variables, that contribute to the overall score generated by a predictive statistical model are presented. In a specialized case, for each variable in the model, an importance is calculated based on the calculated slope and deviance of the predictive variable. In a more general case, for each variable in the model, an importance is calculated based on setting that variable to have the average value for the data set, and then calculating the change in score. The totality of variables (or combinations thereof) is then ranked by the Δscore, or a magnitude of it, such as |Δscore|.

The invention discloses a loaner credit rating method and system based on machine learning. The method herein includes the following steps: acquiring original data for modeling, wherein the original data for modeling includes a credit report and an overdue shop list; extracting from the credit report and classifying indexes, obtaining a predication variable and the weight thereof; on the basis of the overdue shop list, the obtained prediction variable and the weight thereof, using the machine learning method to model, obtaining a prediction model of a response variable and the prediction variable; based on the obtained prediction model, predicting a new loaner, obtaining a default probability of the new loaner; based on the default probability of the new loaner, computing the credit rating of the new loaner. According to the invention, the method herein uses the machine learning method to model, is suitable for meeting fast changing requirements for loaner data under new circumstances; is provided with a step for extracting from the credit report and classifying indexes, therefore being more comprehensive and easier. According to the invention, the method can be widely applied to the field of computer applications.

The invention discloses a demanded power predication based variable-area optimal energy management control method and device, and belongs to the field of hybrid power off-road vehicle energy management. The method comprises the following steps: selecting longitudinal vehicle speed and longitudinal accelerated speed as predicating variables by designing a self-adaptive Markov chain forecasting method, thereby realizing indirect predication of demanded power; controlling the vehicle to work in two different drive modes by calculating and comparing equivalent energy consumption of a hybrid drivemode with that of a pure electric mode in real time; taking the equivalent fuel oil consumption minimum control strategy as developing basis, considering traveling working condition characteristics and demand characteristics of the off-road vehicle, performing real-time solving of an optimizing area according to a demand powder variable, and designing a variable-area equivalent fuel oil consumption minimum control strategy. The demanded powder predication method can improve demanded powder predication precision and self-adaption as well as power system response characteristics, reasonably switches a working mode, optimizes general energy consumption characteristics of the system, and improves travelling stability while guaranteeing working stability of the power system.

The invention discloses a credit monitoring and scoring method and system based on a behavior model. The method comprises the following steps: obtaining the original data of an account in an observation period; carrying out feature extraction on the original data of the account to obtain a prediction variable; according to the prediction variables, using the machine learning algorithm for modeling, and obtaining the prediction model of the response variables and the prediction variables; predicting an account with repayment performance according to the predicting model to obtain a default probability in the predicting period of the account; calculating the behavior score of the account according to the default probability of the account during the prediction period. The invention can increase the data changing speed, improve the applicability and generalization, and carry out accurate behavior scoring on the account.

The invention provides a vehicle spare part sales volume forecasting method and system based on a unified dynamic integration model and a meta-heuristic algorithm. The method comprises the steps thata database is established to store data needed for forecasting the vehicle spare part sales volume, and the sales volume of various vehicle spare parts is comprised and is called as a forecasting variable; a data acquisition module is connected with the database and the vehicle spare part sales volume forecasting system to acquire the needed forecasting variable, and a number of parallel typical forecasting methods are used for forecasting to acquire forecasting results corresponding to various forecasting methods; furthermore, various forecasting results are stored, and a unified dynamic integrated model is established; the meta-heuristic algorithm is used to optimize the forecasting model coefficients; the acquired forecasting model is stored in a vehicle spare part sales volume forecasting application system; and a spare part sales volume forecasting result is generated after the corresponding vehicle spare part sales volume data are input. According to the invention, the model which is suitable for forecasting various vehicle spare parts is found; the characteristics of high optimization precision and the like of the meta-heuristic algorithm are used; and the vehicle spare partsales volume forecasting precision is effectively improved.

A method of modeling includes quantifying a co-operative strength value for a plurality of pairs of variables, and identifying a clique of at least three variables based on a graph of the co-operative strength values of a plurality of pairs of variables. The method also includes selecting a first pair of variables of the plurality of pairs of variables having a high co-operative strength value. A second clique may also be identified. A model of the first clique and a model of the second clique are made. The outputs of these models are combined to form a combined model which is used to make various decisions with respect to real time data.

The invention discloses a fatigue driving prediction method based on GPS data. The fatigue driving prediction method comprises the following steps that 1, a GPS database is built; 2, driving data is selected in the GPS database, the driving data is set as a plurality of prediction variables related with fatigue driving, and the prediction variables are standardized to serve as samples; a fatigue driving situation corresponds to each sample, and the samples are divided into the samples which belong to the fatigue driving state and the samples which do not belong to the fatigue driving state; the samples are divided into a training set and a verification set in a proportion of 7:3; 3, a multi-layer perception neural network including an input layer, a hidden layer and an output layer is built, and a classification model used for predicting fatigue driving is generated; 4, whether the fatigue driving situation can occur or not during the prediction period is predicted by using the model,and when the fatigue driving situation can occur during the prediction period, a driver is reminded. By means of the fatigue driving prediction method, whether the fatigue driving situation occurs ornot can be predicted in advance, and the driver is early warned.

The invention relates to a data prediction model adjusting and optimizing method based on an LSTM network. The method include steps of preprocessing: obtaining previous N small period values of a to-be-predicted variable, and extracting data of previous multiple variables whose correlation coefficient sums are greater than a coefficient threshold in a data set to form a training set; and model training: performing N rounds of training according to a sequence of the period values from large to small to calculate an optimal solution model, wherein each round of training comprises converting thetraining set from time sequence data to a supervised learning sequence; inputting the supervised learning sequence to the LSTM network to obtain a training model of the round; and comparing a root-mean-square error obtained by the training model of the round with a root-mean-square error obtained by training of the previous round, and reserving the training model corresponding to a smaller value as the optimal solution model. The invention also relates to a data prediction model adjusting and optimizing apparatus based on the LSTM network. According to the adjusting and optimizing method and the apparatus, optimization is performed based on the LSTM network, data prediction can be realized, the calculating speed is fast, and the prediction effect is good.

The invention provides a model prediction control method and device for a building air conditioner and relates to the field of control over indoor environments. The model prediction control method comprises the steps that historical thermal environment data of control areas are collected; thermal environments of the control areas in the preset prediction period are predicted according to the historical thermal environment data and predictive variables of the control areas; and according to the prediction result and the time-share energy price, end equipment of the building air conditioner is controlled, so that the control areas are in the comfort degree interval in the preset prediction period, and the energy cost requirement of users are met. According to the technical scheme of the model prediction control method and device for the building air conditioner, optimizing control over the end equipment of the air conditioner is achieved, the comfort degree requirement and the energy cost requirement of the users are considered, the application scene meeting the personalized target requirements of the users is achieved, and the innovative control method that a project is easy to implement, and precision is guaranteed is formed.

The invention discloses a method for predicting a terrorist attack based on the stochastic subspace. The method comprises a first step of establishing a training set, and maintaining a linear classifier Wt by an online learner; a second step of counting the number z of terrorist attack events happening in a country in the next month of the month in a GTD database; step 3: randomly selecting s groups of feature subsets from overall features of a given terrorist attack data set using a stochastic subspace method and generating s base classifiers in an integrated classification algorithm of a kernel extreme learning machine; a fourth step of putting the S groups of feature subsets into the kernel extreme learning machine for learning to obtain output results; a fifth step of integrating the outputs of the s base classifiers to obtain a final classification result; and a sixth step of performing model application: inputting the value of an independent variable for each record in a test set to obtain the value of a predictive variable, that is, the probability of a terrorist attack event happening in the next month.

The invention discloses a frequency device compensation method, device and system and a computer readable storage medium, and belongs to the technical field of communication. The method comprises acquiring a clock signal output by a frequency compensation sensor as a first predictive variable; acquiring a variable output by a predictive variable sensor and related to frequency fluctuation as a second predictive variable; compensating a frequency device according to the first predictive variable and the second predictive variable. By replacing a temperature sensor by a frequency sensor, and providing the predictive variable associated with the frequency fluctuation, the method improves the accuracy and stability of frequency compensation.

The invention discloses a method of detecting artifacts in ICU patient records based on multivariate logistic regression. The method includes: calculating ICU probability of a patient by a multivariate logistic regression expression, and comparing the probability to a set threshold to establishing a binary classifier used to judge whether the ICU patient is going to die or not. The method has the advantages that presetting preconditions having predictive variables obeying normal distribution is not required; the predictive variables can be continuous or disperse; the number of predictive variables in a model is reduced as far as possible, and complexity of the model is reduced while prediction accuracy is guaranteed. The invention further discloses a system to implement the method. The system has the advantages that the problem of non-specificity in existing risk prediction models can be solved effectively; compared to existing models, the model allows prediction to be more accurate and specific.

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 discloses a road traffic accident form prediction method. The method comprises the following steps of: 1, collecting and processing road traffic accident data; 2, discretizing continuous independent variables in the traffic accident data by adopting a minimum description length criterion; 3, mining interaction among the independent variables by adopting an attribute selection method based on association rules in the field of data mining; 4, establishing a hybrid Logit model, and performing parameter estimation by adopting a maximum likelihood estimation method; and 5, based on the constructed hybrid Logit model, predicting the traffic accident form probability. In the continuous independent variable discretization process, information of predictive variables is fully utilized, the influence of the interaction among the variables on the accident form is mined, the information loss of discretized variables is reduced, the problem of wrong inference caused by neglect of the interaction among the variables is solved, the prediction precision of a traffic accident form prediction model is improved, and technical support is provided for improvement of a road traffic safety environment.

The invention provides an econometrics and heuristic intelligence combined automobile sales prediction method and system. The method comprises the following steps of: storing data required by automobile sales prediction through establishing a database, wherein data comprises economic indexes, brand automobile sales and automobile sales, which are called as prediction variables; connecting the database and the automobile sales prediction system through the data so as to obtain required prediction variables, and verifying structural relationships between the variables so as to obtain endogenousvariables with a long-term equilibrium and causal relationship; establishing a vector error correction model for the endogenous variables; optimizing coefficient of a prediction model by utilizing a heuristic intelligence algorithm; and finally storing the obtained prediction model into a sales prediction application system, and generating a sales prediction result after inputting corresponding economic variables, brand automobile sales data and former-year automobile sales data. According to the method and system, a model suitable for long-term prediction is found, and the characteristic of high precision of the heuristic intelligence can be utilized at the same time, so that the automobile sales prediction precision is effectively improved.

A computer-implemented method for prescriptive time-series forecasting, which combines both what-if analysis and goal-seeking analysis. The method comprises building a model for a target metric with a set of predictors, based on historical time-series data, and computing, using the model, a set of forecast values. Using the set of forecast values with respect to a forecasting period, both a set of goals for the target metric and a set of constraints for the predictors are analyzed. A set of updated forecasts based on the analyses with respect to the forecasting period is determined to meet the goals within the set of constraints. The updated set of forecasts is presented with respect to the forecasting period, e.g., using a table, a visualization, and/or an interactive user interface.

The invention discloses a mathematical model for evaluating fertilization abilities of landrace boars and an establishing method for the mathematical model. The establishing method comprises the steps of: firstly, performing data description on predictive variables of data samples of the landrace boars; secondly, performing system clustering, and dividing the landrace boars with different fertilization abilities according to sample data related with fertilization abilities of boar seminal fluids into a boar group with high fertilization ability and a boar group with low fertilization ability; and thirdly, performing discriminant analysis and Logistic regression analysis after clustering to generate a specific mathematical function. The coefficient and prediction accuracy of the mathematical model established by the method are reliable; a function model suitable for evaluating the fertilization abilities of the in-service landrace boars in Guangdong province can be provided; the function model can serve as a basis for evaluating the fertilization abilities of the landrace boars; and the mathematical model is of great significance on reduction of a gradient change interval of the fertilization abilities of the boars and accurate prediction of the fertilization abilities of the landrace boars.

The invention discloses a mathematic model for evaluating the fertility of the Large White, and an establishment method thereof. The establishment method of the mathematic model comprises the following steps: carrying out data description on a prediction variable of a Large White data sample; carrying out system clustering, dividing the Large White of different fertility levels into a Large White group with the high fertility and the Large White group with the low fertility according to sample data relevant to the fertility of Large White sperm; and carrying out discriminant analysis and Logistic regression analysis after clustering is carried out to generate a specific mathematic function. The mathematic model established by the invention has a reliable coefficient and a reliable prediction accurate rate, can provide a function model suitable for Guangdong province boar farms to evaluate the fertility of the in-service Large White, can be taken as a basis of evaluating the fertility of the Large White and has an important meaning for reducing a gradient change interval of the fertility of the Large White and accurately predicting the fertility of the Large White.