CN106682754A - Event occurrence probability prediction method and device - Google Patents
Event occurrence probability prediction method and device Download PDFInfo
- Publication number
- CN106682754A CN106682754A CN201510744935.5A CN201510744935A CN106682754A CN 106682754 A CN106682754 A CN 106682754A CN 201510744935 A CN201510744935 A CN 201510744935A CN 106682754 A CN106682754 A CN 106682754A
- Authority
- CN
- China
- Prior art keywords
- model
- prediction
- probability
- event
- fit
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention provides an event occurrence probability prediction method and device. The event occurrence probability prediction method comprises the steps of according to prediction index type requirements of each prediction in a preset model set for a predicted event, acquiring prediction indexes corresponding to the prediction index type requirements; applying each prediction model to process the corresponding prediction indexes, and acquiring an output probability corresponding to each prediction model; applying a pre-trained fitting model corresponding to the model set to process the output probability corresponding to each prediction model, and predicting the occurrence probability of the event. According to the event occurrence probability prediction method and device, the accuracy of a prediction result is improved, and thus the effectiveness and the reliability of processing relevant services according to the prediction result are improved.
Description
Technical field
The present invention relates to technical field of data processing, more particularly to a kind of event occurrence rate Forecasting Methodology and device.
Background technology
In many business scenarios, it is often necessary to the probability of happening of all kinds of events in prediction business, so as to according to the generation of prediction
Probability is adjusted correspondingly to related service or operates.For example：In Third-party payment business platform, the account of user is predicted
The probability of happening of number stolen event；Or, in financing business platform, predict that generation of the user to institute's monetary allowance volume refund event is general
Rate；Or, in insurance business platform, predict probability of happening of user's Claims Resolution event etc..
At present, according to the probability of happening of all kinds of events in the single forecast model prediction business of business scenario application, such as,
Insurance business platform, is predicted using logistic regression forecast model to the probability of happening of dependent event, in financing business platform,
Application decision tree forecast model is predicted to the probability of happening of dependent event.
However, the true probability of happening of the business event according to historgraphic data recording, the prediction data before analysis, know mesh
The accuracy of front prediction mode can not reach default standard, reduce the validity for processing related service according to predicting the outcome.
The content of the invention
It is contemplated that at least solving one of technical problem in correlation technique to a certain extent.
For this purpose, first purpose of the present invention is to propose a kind of event occurrence rate Forecasting Methodology, improve what is predicted the outcome
The degree of accuracy, and then improve the validity and reliability for processing related service according to predicting the outcome.
Second object of the present invention is to propose a kind of event occurrence rate prediction meanss.
To achieve these goals, the event occurrence rate Forecasting Methodology of first aspect present invention embodiment, including：According to pre-
If prediction index type demand of every kind of forecast model to predicted events in model group, obtain and the prediction index type demand
Corresponding prediction index；Using the corresponding prediction index of every kind of forecast model process, obtain corresponding with every kind of forecast model defeated
Go out probability；Using training in advance model of fit corresponding with the model group, output corresponding with every kind of forecast model is processed
Probability, predicts the probability of happening of the event.
The event occurrence rate Forecasting Methodology of the embodiment of the present invention, first according to every kind of forecast model in preset model group to prediction
The prediction index type demand of event, obtains prediction index corresponding with the prediction index type demand；Then using every kind of
The corresponding prediction index of forecast model process, obtains output probability corresponding with every kind of forecast model；And then apply training in advance
Model of fit corresponding with the model group, process corresponding with every kind of forecast model output probability, predict the event
Probability of happening.Thus, the degree of accuracy for predicting the outcome is improve, and then improves having according to the process related service that predicts the outcome
Effect property and reliability.
To achieve these goals, the event occurrence rate prediction meanss of second aspect present invention embodiment, including：First obtains
Delivery block, for the prediction index type demand according to every kind of forecast model in preset model group to predicted events, obtains and institute
State the corresponding prediction index of prediction index type demand；Second acquisition module, for corresponding using the process of every kind of forecast model
Prediction index, obtains output probability corresponding with every kind of forecast model；Processing module, for using training in advance with it is described
The corresponding model of fit of model group, processes output probability corresponding with every kind of forecast model, predicts the probability of happening of the event.
The event occurrence rate prediction meanss of the embodiment of the present invention, by the first acquisition module according to every kind of pre- in preset model group
Prediction index type demand of the model to predicted events is surveyed, prediction index corresponding with the prediction index type demand is obtained；
Corresponding prediction index is processed by the every kind of forecast model of the second acquisition module application, is obtained corresponding with every kind of forecast model defeated
Go out probability；By the model of fit corresponding with the model group of processing module application training in advance, process and every kind of prediction mould
The corresponding output probability of type, predicts the probability of happening of the event.Thus, the degree of accuracy for predicting the outcome, Jin Erti are improve
The high validity and reliability for processing related service according to predicting the outcome.
The additional aspect of the present invention and advantage will be set forth in part in the description, and partly will from the following description become bright
It is aobvious, or recognized by the practice of the present invention.
Description of the drawings
The above-mentioned and/or additional aspect of the present invention and advantage will be apparent from from the following description of the accompanying drawings of embodiments and
It is easy to understand, wherein,
Fig. 1 is the flow chart of the event occurrence rate Forecasting Methodology of one embodiment of the invention；
Fig. 2 is the flow chart of fitting coefficient training process；
Fig. 3 is the schematic diagram of fitting coefficient training process；
Fig. 4 is the structured flowchart of the event occurrence rate prediction meanss of one embodiment of the invention；
Fig. 5 is the structured flowchart of the event occurrence rate prediction meanss of another embodiment of the present invention.
Specific embodiment
Embodiments of the invention are described below in detail, the example of the embodiment is shown in the drawings, wherein identical from start to finish
Or similar label represents same or similar element or the element with same or like function.Retouch below with reference to accompanying drawing
The embodiment stated is exemplary, is only used for explaining the present invention, and is not considered as limiting the invention.Conversely, this
Bright embodiment includes all changes, modification and the equivalent fallen into the range of the spirit and intension of attached claims.
In describing the invention, it is to be understood that term " first ", " second " etc. are only used for describing purpose, and can not manage
Solve to indicate or implying relative importance.In describing the invention, it should be noted that unless otherwise clearly regulation and
Limit, term " connected ", " connection " should be interpreted broadly, for example, it may be fixedly connected, or be detachably connected,
Or be integrally connected；Can be mechanically connected, or electrically connect；Can be joined directly together, it is also possible to by middle matchmaker
Jie is indirectly connected to.For the ordinary skill in the art, can understand above-mentioned term in the present invention with concrete condition
Concrete meaning.Additionally, in describing the invention, unless otherwise stated, " multiple " are meant that two or more.
In flow chart or here any process described otherwise above or method description are construed as, expression includes one
Or more module, fragment or parts for being used for the code of executable instruction the step of realize specific logical function or process,
And the scope of the preferred embodiment of the present invention includes other realization, wherein order that is shown or discussing can not be pressed,
Including according to involved function by it is basic simultaneously in the way of or in the opposite order, carry out perform function, this should be by the present invention's
Embodiment person of ordinary skill in the field understood.
Below in conjunction with Description of Drawings event occurrence rate Forecasting Methodology according to embodiments of the present invention and device.
Fig. 1 is the flow chart of the event occurrence rate Forecasting Methodology of one embodiment of the invention.
As shown in figure 1, the event occurrence rate Forecasting Methodology includes：
Step 101, according to prediction index type demand of the every kind of forecast model in preset model group to predicted events, obtains and institute
State the corresponding prediction index of prediction index type demand.
Specifically, event occurrence rate Forecasting Methodology provided in an embodiment of the present invention is configured in event occurrence rate prediction meanss
In, wherein, event occurrence rate prediction meanss can be deployed in the business platform of different business scene with integrated, to predict
The probability of happening of all kinds of events in business；Or, event occurrence rate prediction meanss can be disposed individually, flat with each business
Data interaction is carried out by default interface between platform, at the prediction index to the business to be predicted transmitted in business platform
After reason, will predict the outcome and be sent to corresponding business platform by interface.
Model group is previously provided with event occurrence rate prediction meanss, wherein, model group includes at least two forecast models,
The species of forecast model is a lot, can be needed to be selected according to practical application, and the present embodiment is not restricted to this, for example may be used
To include：Logic Regression Models, C5.0 decision-tree models, neural network model, prob regression models and CART are returned
Tree-model.
Because business scenario is various, related miscellaneous service event content is also different, therefore, predict that the generation of different event is general
Prediction index type content required for rate is different.Also, when being predicted for same part event, different forecast model institutes
The input prediction pointer type and specific algorithm of needs is also different.Therefore, when the probability of happening to concrete business event enters
During row prediction, the related prediction index of probability event is derived according to service needed, and extracted comprising history index sample
Collection and target variable are (generally, if event there occurs that target variable is 1, without the historical sample for occurring 0).Need
Illustrate, the acquisition modes of prediction index are a lot, according to specific application scenarios depending on, for example：Can prestore
It is in event occurrence rate prediction meanss, or by the interface between associated traffic data storehouse from Service Database
In in real time obtain.
For the acquisition process for more clearly illustrating prediction index, included with model group：Logic Regression Models and decision tree
Two kinds of forecast models of model, for Alipay platform, as a example by predicting the probability of happening of the stolen event of Alipay account of user,
It is described as follows：
When being predicted using Logic Regression Models, need (to be designated as history index sample using the historical sample comprising predictive variable
Collection A1) prediction user the stolen event of Alipay account probability of happening pred_1, note model be mode1；Application decision tree
When model is predicted, the probability of happening that the stolen event of precious account is got paid according to history index sample set A1 fittings is needed
Pred_2, note model is model2；The input pointer for assuming decision-tree model and logistic regression mould adds up to 30, therefore,
According to the prediction index type demand of decision-tree model and logistic regression mould to predicted events, obtain defeated with 30 kinds from database
Enter the corresponding prediction index of index.Wherein, prediction index type is included a lot, such as：User account, current balance,
Pay list, time of opening an account, record of bad behavior, credit rating, accrediting amount etc. in detail.
Step 102, using the corresponding prediction index of every kind of forecast model process, obtains output corresponding with every kind of forecast model general
Rate.
Specifically, obtain after prediction index corresponding with the index demand of every kind of forecast model, using the process of every kind of forecast model
Corresponding prediction index, obtains output probability corresponding with every kind of forecast model.The different algorithm of different forecast model applications
Corresponding prediction index is processed, so as to obtain output probability corresponding with every kind of forecast model.
Step 103, using training in advance model of fit corresponding with the model group, is processed corresponding with every kind of forecast model
Output probability, predicts the probability of happening of the event.
Specifically, the good model of fit corresponding with the model group of training in advance is provided with event occurrence rate prediction meanss,
Wherein, model of fit is to being fitted prediction index type and fitting coefficient combination producing, fitting therein by logical operator
Prediction index type is exactly the output probability of each forecast model in model group, and fitting coefficient is to instruction previously according to object function
Practice what sample set training determined, wherein, the fitting coefficient, including：It is corresponding with every kind of forecast model in the model group
Coefficient, and/or；Constant coefficient corresponding with the model of fit.
It should be noted that many to being fitted prediction index type and fitting coefficient combination by logical operator, can be with
According to being configured using needs, the present embodiment is not restricted to this.
After obtaining output probability corresponding with every kind of forecast model, using training in advance model of fit corresponding with model group,
Output probability corresponding with every kind of forecast model is processed, the probability of happening of the event is predicted.
In order to more clearly illustrate the prediction process to event occurrence rate, continue to be illustrated with the example of step 101：
Model group includes：Two kinds of forecast models of Logic Regression Models model1 and decision-tree model model2, using model1 and
Predict the outcome and the logit regression algorithms of model2 is fitted the final mask model3 for obtaining：
Pred=1/ (1+exp (- Z)), wherein, Z=a1*pred_1+a2*pred_2+a0；
Wherein, pred_1 is output probability corresponding with Logic Regression Models model1；Pred_2 is and decision-tree model
The corresponding output probabilities of model2；A1, a2 and a0 are fitting coefficient.
The probability of happening of the event is predicted using model3:Exported using model1 and the prediction index into model
Probability p red_1；Output probability pred_2 is obtained using model2 and into the prediction index of model；By pred_1 and pred_2
Bring into the formula of model3 and obtain the probability of happening of the event.
The event occurrence rate Forecasting Methodology of the present embodiment, first according to finger of every kind of forecast model to predicted events in model group
Mark demand, obtains prediction index corresponding with the index demand；Then refer to using the corresponding prediction of every kind of forecast model process
Mark, obtains output probability corresponding with every kind of forecast model；And then using the fitting corresponding with the model group of training in advance
Model, processes output probability corresponding with every kind of forecast model, predicts the probability of happening of the event.Thus, improve pre-
The degree of accuracy of result is surveyed, and then improves the validity and reliability for processing related service according to predicting the outcome.
Based on above-described embodiment, for the step 103 shown in Fig. 1 before, need to carry out the fitting coefficient in model of fit
Training, below by Fig. 2 and Fig. 3 the training process of fitting coefficient is illustrated, specific as follows：
Fig. 2 is the flow chart of fitting coefficient training process, and Fig. 3 is the schematic diagram of fitting coefficient training process,
As shown in Figures 2 and 3, the specific training process of fitting coefficient may comprise steps of：
Step 201, obtains the history index sample set for being used for training the model group, and corresponding with the predicted events
History probability of happening sample set.
Step 202, obtains history corresponding with every kind of forecast model in the model group defeated according to the history index sample set
Go out probability sample collection.
Step 203, using the history output probability sample set as the model of fit input, and by the history send out
Raw probability sample collection determines the fitting coefficient in the model of fit as the output of the model of fit.
Specifically, interface is set between event occurrence rate prediction meanss and training sample set database, wherein, training sample
Collection database can be Service Database, or, the third-party platform of storage service record.By interface and training sample set
Connection is set up between database, from training sample set database the history index sample set A1 for training pattern group is obtained, with
And history probability of happening sample set corresponding with predicted events, according to history index sample set A1 obtain with the model group in
The corresponding history output probability sample set A2 of every kind of forecast model, using history output probability sample set A2 as with model group pair
The input of the model of fit answered, and train history output general as the output of model of fit history probability of happening sample set
Rate sample set A2 and history probability of happening sample set are consistent, so that it is determined that the fitting coefficient in model of fit.
Referring to Fig. 3, and with continued reference to the example in step 103, according to multiple in history index sample set A1 and model group
Forecast model obtains history output probability sample set A2 corresponding with every kind of forecast model, using logit regression algorithms
Pred=1/ (1+exp (- Z)), wherein, Z=a1*pred_1+a2*pred_2+a0, usage history output probability sample set A2
That is the predictive variable that pred_1 ... pred_n is input into as model of fit, usage history probability of happening sample set is used as model of fit
Output, fitting coefficient a1, a2 and a0 are determined, so as to obtain final model of fit.
The event occurrence rate Forecasting Methodology of the present embodiment, by obtaining the history index sample set for being used for training the model group,
And history probability of happening sample set corresponding with the predicted events, obtained and the mould according to the history index sample set
The corresponding history output probability sample set of every kind of forecast model in type group, using the history output probability sample set as the plan
The input of matched moulds type, and the history probability of happening sample set is determined into the fitting as the output of the model of fit
Fitting coefficient in model.Thus, the degree of accuracy for predicting the outcome is improve, and then is improve related according to the process that predicts the outcome
The validity and reliability of business.
In order to realize above-described embodiment, embodiments of the invention also provide a kind of event occurrence rate prediction meanss.
Fig. 4 is the structured flowchart of event occurrence rate prediction meanss according to an embodiment of the invention.
As shown in figure 4, the event occurrence rate prediction meanss include：
First acquisition module 11, for the prediction index type need according to every kind of forecast model in preset model group to predicted events
Ask, obtain prediction index corresponding with the prediction index type demand；
Second acquisition module 12, for using the corresponding prediction index of every kind of forecast model process, obtaining and every kind of forecast model
Corresponding output probability；
Processing module 13, for using training in advance model of fit corresponding with the model group, processing and every kind of prediction mould
The corresponding output probability of type, predicts the probability of happening of the event.
Wherein, the model group includes following at least two forecast model：
Logic Regression Models, decision-tree model, neural network model, prob regression models and CART regression tree models.
In one embodiment, if the model group includes：Logic Regression Models model 1 and decision-tree model model2 two
Plant forecast model,
The processing module 13 specifically for：Using the model of fit model3 obtained beforehand through logistic regression Algorithm for Training
Output probability corresponding with every kind of forecast model is processed, the probability of happening of the event is predicted, wherein, the model3 is：
Pred=1/ (1+exp (- Z)), wherein, Z=a1*pred_1+a2*pred_2+a0；
Wherein, pred_1 is and the corresponding output probability of Logic Regression Models model 1；Pred_2 is and decision-tree model
The corresponding output probabilities of model2；Pred is the probability of happening of the event；A1, a2 and a0 are fitting coefficient.
Wherein, in one embodiment, first acquisition module 11, specifically for：
Obtain prediction corresponding with the prediction index type demand from Service Database corresponding with the predicted events to refer to
Mark.
It should be noted that the aforementioned explanation to event occurrence rate Forecasting Methodology embodiment is also applied for the embodiment
Event occurrence rate prediction meanss, here is omitted.
The event occurrence rate prediction meanss of the embodiment of the present invention, first according to every kind of forecast model in preset model group to prediction
The prediction index type demand of event, obtains prediction index corresponding with the prediction index type demand；Then using every kind of
The corresponding prediction index of forecast model process, obtains output probability corresponding with every kind of forecast model；And then apply training in advance
Model of fit corresponding with the model group, process corresponding with every kind of forecast model output probability, predict the event
Probability of happening.Thus, the degree of accuracy for predicting the outcome is improve, and then improves having according to the process related service that predicts the outcome
Effect property and reliability.
Fig. 5 is the structured flowchart of the event occurrence rate prediction meanss of another embodiment of the present invention.
As shown in figure 5, being based on Fig. 4 embodiments, also include：
3rd acquisition module 14, is used for training the history index sample set of the model group for obtaining, and with the prediction
The corresponding history probability of happening sample set of event.
Second acquisition module 12, is additionally operable to be obtained and every kind of forecast model in the model group according to the history index sample set
Corresponding history output probability sample set
Training module 15, for using the history output probability sample set as the model of fit input, and will be described
History probability of happening sample set determines the fitting coefficient in the model of fit as the output of the model of fit.
Wherein, the fitting coefficient, including：
Coefficient corresponding with every kind of forecast model in the model group, and/or；
Constant coefficient corresponding with the model of fit.
It should be noted that the aforementioned explanation to event occurrence rate Forecasting Methodology embodiment is also applied for the embodiment
Event occurrence rate prediction meanss, here is omitted.
The event occurrence rate prediction meanss of the embodiment of the present invention, by obtaining the history index sample for being used for training the model group
This collection, and history probability of happening sample set corresponding with the predicted events, according to the history index sample set obtain with
The corresponding history output probability sample set of every kind of forecast model in the model group, using the history output probability sample set as
The input of the model of fit, and the history probability of happening sample set is determined into institute as the output of the model of fit
State the fitting coefficient in model of fit.Thus, the degree of accuracy for predicting the outcome is improve, and then is improve according to the place that predicts the outcome
The validity and reliability of reason related service.
It should be appreciated that each several part of the present invention can be realized with hardware, software, firmware or combinations thereof.In above-mentioned reality
In applying mode, software that multiple steps or method can in memory and by suitable instruction execution system be performed with storage or
Firmware is realizing.For example, if realized with hardware, and in another embodiment, can be with well known in the art
Any one of row technology or their combination are realizing：With for realizing the logic gates of logic function to data-signal
Discrete logic, the special IC with suitable combinational logic gate circuit, programmable gate array (PGA) is existing
Field programmable gate array (FPGA) etc..
In the description of this specification, reference term " one embodiment ", " some embodiments ", " example ", " specific example ",
Or the description of " some examples " etc. means to combine specific features, structure, material or feature bag that the embodiment or example are described
In being contained at least one embodiment of the present invention or example.In this manual, to the schematic representation of above-mentioned term not necessarily
Refer to identical embodiment or example.And, the specific features of description, structure, material or feature can be any
Combine in an appropriate manner in one or more embodiments or example.
Although an embodiment of the present invention has been shown and described, it will be understood by those skilled in the art that：Without departing from
In the case of the principle and objective of the present invention various changes, modification, replacement and modification can be carried out to these embodiments, this
Bright scope is limited by claim and its equivalent.
Claims (10)
1. a kind of event occurrence rate Forecasting Methodology, it is characterised in that comprise the following steps：
According to prediction index type demand of the every kind of forecast model in preset model group to predicted events, obtain and refer to the prediction
The corresponding prediction index of mark type demand；
Using the corresponding prediction index of every kind of forecast model process, output probability corresponding with every kind of forecast model is obtained；
Using training in advance model of fit corresponding with the model group, output probability corresponding with every kind of forecast model is processed,
Predict the probability of happening of the event.
2. event occurrence rate Forecasting Methodology as claimed in claim 1, it is characterised in that if the model group includes：Patrol
Two kinds of forecast models of regression model model1 and decision-tree model model2 are collected,
The model of fit corresponding with the model group of the application training in advance, processes output corresponding with every kind of forecast model
Probability, predicting the probability of happening of the event includes：
Process corresponding with every kind of forecast model defeated using the model of fit model3 obtained beforehand through logistic regression Algorithm for Training
Go out probability, predict the probability of happening of the event, wherein, the model3 is：
Pred=1/ (1+exp (- Z)), wherein, Z=a1*pred_1+a2*pred_2+a0；
Wherein, pred_1 is output probability corresponding with Logic Regression Models model1；Pred_2 is and decision-tree model
The corresponding output probabilities of model2；Pred is the probability of happening of the event；A1, a2 and a0 are fitting coefficient.
3. event occurrence rate Forecasting Methodology as claimed in claim 1, it is characterised in that the acquisition refers to the prediction
The corresponding prediction index of mark type demand, including：
Obtain prediction corresponding with the prediction index type demand from Service Database corresponding with the predicted events to refer to
Mark.
4. the event occurrence rate Forecasting Methodology as described in claim 1-4 is arbitrary, it is characterised in that also include：
The history index sample set for being used for training the model group is obtained, and history corresponding with the predicted events occurs generally
Rate sample set；
History output probability sample corresponding with every kind of forecast model in the model group is obtained according to the history index sample set
Collection；
Using the history output probability sample set as the model of fit input, and by the history probability of happening sample
Collect the output as the model of fit, determine the fitting coefficient in the model of fit.
5. event occurrence rate Forecasting Methodology as claimed in claim 4, it is characterised in that the fitting coefficient, including：
Coefficient corresponding with every kind of forecast model in the model group, and/or；
Constant coefficient corresponding with the model of fit.
6. a kind of event occurrence rate prediction meanss, it is characterised in that include：
First acquisition module, for the prediction index type demand according to every kind of forecast model in preset model group to predicted events,
Obtain prediction index corresponding with the prediction index type demand；
Second acquisition module, for using the corresponding prediction index of every kind of forecast model process, obtaining and every kind of forecast model pair
The output probability answered；
Processing module, for using training in advance model of fit corresponding with the model group, processing and every kind of forecast model
Corresponding output probability, predicts the probability of happening of the event.
7. event occurrence rate prediction meanss as claimed in claim 6, it is characterised in that if the model group includes：Patrol
Two kinds of forecast models of regression model model1 and decision-tree model model2 are collected,
The processing module specifically for：Using the model of fit model3 process obtained beforehand through logistic regression Algorithm for Training
Output probability corresponding with every kind of forecast model, predicts the probability of happening of the event, wherein, the model3 is：
Pred=1/ (1+exp (- Z)), wherein, Z=a1*pred_1+a2*pred_2+a0；
Wherein, pred_1 is output probability corresponding with Logic Regression Models model1；Pred_2 is and decision-tree model
The corresponding output probabilities of model2；Pred is the probability of happening of the event；A1, a2 and a0 are fitting coefficient.
8. event occurrence rate prediction meanss as claimed in claim 6, it is characterised in that first acquisition module, tool
Body is used for：
Obtain prediction corresponding with the prediction index type demand from Service Database corresponding with the predicted events to refer to
Mark.
9. event occurrence rate prediction meanss as described in claim 6-8 is arbitrary, it is characterised in that also include：
3rd acquisition module, is used for training the history index sample set of the model group for obtaining, and with the prediction thing
The corresponding history probability of happening sample set of part；
Second acquisition module, is additionally operable to be obtained and every kind of prediction mould in the model group according to the history index sample set
The corresponding history output probability sample set of type；
Training module, as the input of the model of fit, and goes through for using the history output probability sample set by described
History probability of happening sample set determines the fitting coefficient in the model of fit as the output of the model of fit.
10. event occurrence rate prediction meanss as claimed in claim 9, it is characterised in that the fitting coefficient, including：
Coefficient corresponding with every kind of forecast model in the model group, and/or；
Constant coefficient corresponding with the model of fit.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510744935.5A CN106682754A (en) | 2015-11-05 | 2015-11-05 | Event occurrence probability prediction method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510744935.5A CN106682754A (en) | 2015-11-05 | 2015-11-05 | Event occurrence probability prediction method and device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN106682754A true CN106682754A (en) | 2017-05-17 |
Family
ID=58858783
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510744935.5A Pending CN106682754A (en) | 2015-11-05 | 2015-11-05 | Event occurrence probability prediction method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106682754A (en) |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107832893A (en) * | 2017-11-24 | 2018-03-23 | 广东电网有限责任公司电力科学研究院 | Power transmission and transforming equipment damage probability forecasting method and device under typhoon based on logistic |
| CN108282586A (en) * | 2017-07-26 | 2018-07-13 | 平安科技（深圳）有限公司 | Prediction technique, device, storage medium based on service line and terminal |
| CN110245237A (en) * | 2018-03-09 | 2019-09-17 | 北京国双科技有限公司 | Event prediction method and device |
| CN110458328A (en) * | 2019-07-11 | 2019-11-15 | 清华大学 | Black Swan event decision method and device based on subjective and objective associated prediction |
| CN111105082A (en) * | 2019-12-05 | 2020-05-05 | 山东浪潮人工智能研究院有限公司 | Workpiece quality prediction model construction method and prediction method based on machine learning |
| CN111192071A (en) * | 2018-11-15 | 2020-05-22 | 北京嘀嘀无限科技发展有限公司 | Invoice amount estimation method and device and invoice probability model training method and device |
| CN111626783A (en) * | 2020-04-30 | 2020-09-04 | 贝壳技术有限公司 | Offline information setting method and device for realizing event conversion probability prediction |
| CN112381338A (en) * | 2021-01-14 | 2021-02-19 | 北京新唐思创教育科技有限公司 | Event probability prediction model training method, event probability prediction method and related device |
| CN112598500A (en) * | 2020-12-21 | 2021-04-02 | 中国建设银行股份有限公司 | Credit processing method and system for non-limit client |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103679287A (en) * | 2013-12-05 | 2014-03-26 | 王海燕 | Combined type power load forecasting method |
| CN103903071A (en) * | 2014-04-17 | 2014-07-02 | 上海电机学院 | Wind power forecast combination method and system |
| CN104182474A (en) * | 2014-07-30 | 2014-12-03 | 北京拓明科技有限公司 | Method for recognizing pre-churn users |
| CN104900061A (en) * | 2015-05-29 | 2015-09-09 | 内蒙古工业大学 | Road section travel time monitoring method and device |
-
2015
- 2015-11-05 CN CN201510744935.5A patent/CN106682754A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103679287A (en) * | 2013-12-05 | 2014-03-26 | 王海燕 | Combined type power load forecasting method |
| CN103903071A (en) * | 2014-04-17 | 2014-07-02 | 上海电机学院 | Wind power forecast combination method and system |
| CN104182474A (en) * | 2014-07-30 | 2014-12-03 | 北京拓明科技有限公司 | Method for recognizing pre-churn users |
| CN104900061A (en) * | 2015-05-29 | 2015-09-09 | 内蒙古工业大学 | Road section travel time monitoring method and device |
Cited By (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108282586A (en) * | 2017-07-26 | 2018-07-13 | 平安科技（深圳）有限公司 | Prediction technique, device, storage medium based on service line and terminal |
| CN108282586B (en) * | 2017-07-26 | 2019-02-19 | 平安科技（深圳）有限公司 | Prediction technique, device, storage medium and terminal based on service line |
| CN107832893A (en) * | 2017-11-24 | 2018-03-23 | 广东电网有限责任公司电力科学研究院 | Power transmission and transforming equipment damage probability forecasting method and device under typhoon based on logistic |
| CN110245237A (en) * | 2018-03-09 | 2019-09-17 | 北京国双科技有限公司 | Event prediction method and device |
| CN111192071A (en) * | 2018-11-15 | 2020-05-22 | 北京嘀嘀无限科技发展有限公司 | Invoice amount estimation method and device and invoice probability model training method and device |
| CN111192071B (en) * | 2018-11-15 | 2023-11-17 | 北京嘀嘀无限科技发展有限公司 | Method and device for estimating amount of bill, method and device for training bill probability model |
| CN110458328A (en) * | 2019-07-11 | 2019-11-15 | 清华大学 | Black Swan event decision method and device based on subjective and objective associated prediction |
| CN111105082A (en) * | 2019-12-05 | 2020-05-05 | 山东浪潮人工智能研究院有限公司 | Workpiece quality prediction model construction method and prediction method based on machine learning |
| CN111626783A (en) * | 2020-04-30 | 2020-09-04 | 贝壳技术有限公司 | Offline information setting method and device for realizing event conversion probability prediction |
| CN111626783B (en) * | 2020-04-30 | 2021-08-31 | 贝壳找房(北京)科技有限公司 | Offline information setting method and device for realizing event conversion probability prediction |
| CN112598500A (en) * | 2020-12-21 | 2021-04-02 | 中国建设银行股份有限公司 | Credit processing method and system for non-limit client |
| CN112381338A (en) * | 2021-01-14 | 2021-02-19 | 北京新唐思创教育科技有限公司 | Event probability prediction model training method, event probability prediction method and related device |
| CN112381338B (en) * | 2021-01-14 | 2021-07-27 | 北京新唐思创教育科技有限公司 | Event probability prediction model training method, event probability prediction method and related device |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106682754A (en) | Event occurrence probability prediction method and device | |
| DeStefano et al. | Method for estimating transition probability in bridge deterioration models | |
| CN110222880B (en) | Service risk determining method, model training method and data processing method | |
| CN110830448B (en) | Target event flow abnormity detection method and device, electronic equipment and medium | |
| CN105631600A (en) | Method and system for publishing auto insurance survey tasks | |
| CN108681961A (en) | Credit product promotion method, apparatus, equipment and computer readable storage medium | |
| CN109255517B (en) | Method and device for generating wind control strategy, server and readable storage medium | |
| CN108334633A (en) | Data-updating method, device, computer equipment and storage medium | |
| CN109146661A (en) | User type prediction technique, device, electronic equipment and storage medium | |
| CN108900316A (en) | Car networking traffic management platform and server | |
| Kamariotis et al. | A metric for assessing and optimizing data-driven prognostic algorithms for predictive maintenance | |
| CN110458633A (en) | Invoice data generation method, system and computer equipment based on block chain technology | |
| CN104156442A (en) | Method and system for multi-dimension accumulation on basis of dynamic data and static data and accumulated data application method | |
| US20140052431A1 (en) | Supporting proactive decision-making in event-driven applications | |
| Kim et al. | A study toward appropriate architecture of system-level prognostics: physics-based and data-driven approaches | |
| CN116880867A (en) | Policy big model-based decision engine updating method and device | |
| CN110490563A (en) | Data processing method, device, computer equipment and storage medium | |
| CN111583010A (en) | Data processing method, device, equipment and storage medium | |
| CN115545276A (en) | Order receiving rate prediction method and system for abnormal orders of online taxi appointment | |
| CN114943374A (en) | Subway OD (optical density) quantity prediction method, equipment and storage medium | |
| CN109743203B (en) | Distributed service security combination system and method based on quantitative information flow | |
| CN110087230B (en) | Data processing method, data processing device, storage medium and electronic equipment | |
| US20220147954A1 (en) | Prediction of billing disputes | |
| CN115578854B (en) | Method, device, equipment and storage medium for predicting queuing length of toll station vehicle | |
| CN109919682A (en) | A kind of product and technology of utilizing parameter market adaptability automatic analysis system and method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20170517 |