CN114186096A - Information processing method and device - Google Patents
Information processing method and device Download PDFInfo
- Publication number
- CN114186096A CN114186096A CN202111510017.8A CN202111510017A CN114186096A CN 114186096 A CN114186096 A CN 114186096A CN 202111510017 A CN202111510017 A CN 202111510017A CN 114186096 A CN114186096 A CN 114186096A
- Authority
- CN
- China
- Prior art keywords
- variables
- variable
- matrix
- data set
- characteristic data
- 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
- 230000010365 information processing Effects 0.000 title claims abstract description 44
- 238000003672 processing method Methods 0.000 title claims abstract description 36
- 230000001364 causal effect Effects 0.000 claims abstract description 70
- 238000002474 experimental method Methods 0.000 claims abstract description 46
- 239000011159 matrix material Substances 0.000 claims description 234
- 230000006870 function Effects 0.000 claims description 91
- 238000005457 optimization Methods 0.000 claims description 52
- 238000004590 computer program Methods 0.000 claims description 8
- 238000000034 method Methods 0.000 abstract description 7
- 238000013500 data storage Methods 0.000 description 8
- 230000000694 effects Effects 0.000 description 8
- 230000003203 everyday effect Effects 0.000 description 8
- 238000004458 analytical method Methods 0.000 description 7
- 230000002354 daily effect Effects 0.000 description 4
- 238000010586 diagram Methods 0.000 description 4
- 230000014509 gene expression Effects 0.000 description 4
- 230000001419 dependent effect Effects 0.000 description 3
- 238000013459 approach Methods 0.000 description 2
- 230000003190 augmentative effect Effects 0.000 description 2
- 230000005540 biological transmission Effects 0.000 description 2
- 238000007796 conventional method Methods 0.000 description 2
- 238000002790 cross-validation Methods 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 230000006978 adaptation Effects 0.000 description 1
- 238000003491 array Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 239000004973 liquid crystal related substance Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/70—Information retrieval; Database structures therefor; File system structures therefor of video data
- G06F16/73—Querying
- G06F16/735—Filtering based on additional data, e.g. user or group profiles
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Computational Linguistics (AREA)
- Multimedia (AREA)
- Data Mining & Analysis (AREA)
- Databases & Information Systems (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- User Interface Of Digital Computer (AREA)
Abstract
The present disclosure relates to an information processing method and apparatus, the information processing method including: acquiring an experimental data set and an observed characteristic data set, wherein the characteristic data set comprises characteristic data generated by each user account in a plurality of user accounts aiming at each variable in a plurality of variables, the experimental data set is the characteristic data obtained by carrying out experiments aiming at the user accounts and the experimental variables, the experimental variables at least comprise a plurality of variables, and the characteristic data is the operation information of the user accounts on a preset platform; and obtaining a causal relationship among the multiple variables according to the experimental data set and the observed characteristic data set, wherein the causal relationship among the multiple variables represents an association relationship among different operation information. The method and the device can solve the problem of inaccurate inference of the causal association relation between the operation information of the user account caused by missing data in the related art.
Description
Technical Field
The present disclosure relates to the field of data analysis, and in particular, to an information processing method and apparatus.
Background
For causal analysis with more independent variables, it is necessary to know whether there is a relationship between the independent variables and whether a variable has an indirect influence or a direct influence on the dependent variable, which often involves the problem of discrete search. However, this approach may bias the estimation of a larger weight matrix due to the problem of missing variables, making the inference of causal relationships between variables inaccurate. For example, for a network platform, data of a part of variables such as daily idle time length of a user account cannot be observed, and in this case, when a user wishes to watch a video of the network platform in idle time, the recommended video duration of the network platform may be too long or too short due to the inability to observe the data of the daily idle time length of the user account, and the user's requirement may not be met. In order to better meet the requirements of users, inaccurate inference of causal association between operation information of user accounts caused by missing data needs to be solved when causal analysis of data related to a network platform is performed.
Disclosure of Invention
The present disclosure provides an information processing method and apparatus to solve at least the problems of the related art described above, and may not solve any of the problems described above. The technical scheme of the disclosure is as follows:
according to a first aspect of the embodiments of the present disclosure, there is provided an information processing method including: acquiring an experimental data set and an observed characteristic data set, wherein the characteristic data set comprises characteristic data generated by each user account in a plurality of user accounts for each variable in a plurality of variables, the experimental data set is characteristic data obtained by performing experiments on the plurality of user accounts and experimental variables, the experimental variables at least comprise the plurality of variables, and the characteristic data is operation information of the user accounts on a predetermined platform; and obtaining the causal relationship among the variables according to the experimental data set and the observed characteristic data set, wherein the causal relationship among the variables represents the association relationship among different operation information.
Optionally, the obtaining a causal relationship between the plurality of variables according to the experimental data set and the observed characteristic data set includes: deriving an input matrix from said observed feature data set, and deriving a tool variable matrix from said experimental data set, wherein elements of each row of the input matrix are feature data generated by the same user account for each variable of the plurality of variables, the elements of each column of the input matrix are feature data generated for the same variable for each of the plurality of user accounts, the elements of each row of the tool variable matrix are feature data generated by the same user account for each experiment of a plurality of experiments, the elements of each column of the tool variable matrix are feature data generated for each of the plurality of user accounts for the same experiment, each experiment of the plurality of experiments is conducted with respect to at least one of the experimental variables; and obtaining a weight matrix according to the input matrix and the tool variable matrix, wherein the weight matrix represents the causal relationship among the variables.
Optionally, obtaining a weight matrix according to the input matrix and the tool variable matrix includes: and optimizing a predefined objective function according to the input matrix and the tool variable matrix to obtain a weight matrix which enables the value of the predefined objective function to be minimum.
Optionally, before the optimizing a predefined objective function according to the input matrix and the tool variable matrix to obtain a weight matrix that minimizes a value of the predefined objective function, the method further includes: and limiting the values of the elements in the weight matrix according to the predefined relationship among the variables.
Optionally, the predefined objective function comprises an optimization term, the optimization term is a sum of an optimization function of each endogenous variable and an optimization function of each exogenous variable, wherein the endogenous variable and the exogenous variable are variables of the plurality of variables, the endogenous variable is related to an unobserved variable, the exogenous variable is not related to the unobserved variable, and the unobserved variable has an effect on a causal relationship between the plurality of variables and is not a variable of the plurality of variables.
Optionally, for a variable corresponding to an element in an ith column of the input matrix, when the variable corresponding to the element in the ith column is an endogenous variable, an optimization function of the variable corresponding to the element in the ith column is obtained by: obtaining a residual error of the ith column of the input matrix regressing the columns of the input matrix except the ith column according to the ith column of the input matrix, the columns of the input matrix except the ith column and the weight matrix; obtaining a weighting matrix according to the residual error, the tool variable matrix and the number of the user accounts of the plurality of user accounts; and obtaining an optimization function of the variable corresponding to the element of the ith column according to the residual error, the tool variable matrix and the weighting matrix.
Optionally, for a variable corresponding to an element in an ith column of the input matrix, when the variable corresponding to the element in the ith column is an exogenous variable, an optimization function of the variable corresponding to the element in the ith column is obtained by: obtaining a residual error of the ith column of the input matrix regressing the columns of the input matrix except the ith column according to the ith column of the input matrix, the columns of the input matrix except the ith column and the weight matrix; and obtaining an optimization function of the variable corresponding to the element of the ith row according to the residual error.
According to a second aspect of the embodiments of the present disclosure, there is provided an information processing apparatus including: an acquisition unit configured to: acquiring an experimental data set and an observed characteristic data set, wherein the characteristic data set comprises characteristic data generated by each user account in a plurality of user accounts for each variable in a plurality of variables, the experimental data set is characteristic data obtained by performing experiments on the plurality of user accounts and experimental variables, the experimental variables at least comprise the plurality of variables, and the characteristic data is operation information of the user accounts on a predetermined platform; a determination unit configured to: and obtaining the causal relationship among the variables according to the experimental data set and the observed characteristic data set, wherein the causal relationship among the variables represents the association relationship among different operation information.
Optionally, the determining unit is configured to: deriving an input matrix from said observed feature data set, and deriving a tool variable matrix from said experimental data set, wherein elements of each row of the input matrix are feature data generated by the same user account for each variable of the plurality of variables, the elements of each column of the input matrix are feature data generated for the same variable for each of the plurality of user accounts, the elements of each row of the tool variable matrix are feature data generated by the same user account for each experiment of a plurality of experiments, the elements of each column of the tool variable matrix are feature data generated for each of the plurality of user accounts for the same experiment, each experiment of the plurality of experiments is conducted with respect to at least one of the experimental variables; and obtaining a weight matrix according to the input matrix and the tool variable matrix, wherein the weight matrix represents the causal relationship among the variables.
Optionally, the determining unit is configured to: and optimizing a predefined objective function according to the input matrix and the tool variable matrix to obtain a weight matrix which enables the value of the predefined objective function to be minimum.
Optionally, the apparatus further comprises an adjusting unit configured to: and before optimizing a predefined objective function according to the input matrix and the tool variable matrix to obtain a weight matrix which enables the value of the predefined objective function to be minimum, limiting the value of an element in the weight matrix according to the predefined relation among the variables.
Optionally, the predefined objective function comprises an optimization term, the optimization term is a sum of an optimization function of each endogenous variable and an optimization function of each exogenous variable, wherein the endogenous variable and the exogenous variable are variables of the plurality of variables, the endogenous variable is related to an unobserved variable, the exogenous variable is not related to the unobserved variable, and the unobserved variable has an effect on a causal relationship between the plurality of variables and is not a variable of the plurality of variables.
Optionally, for a variable corresponding to an element in an ith column of the input matrix, when the variable corresponding to the element in the ith column is an endogenous variable, an optimization function of the variable corresponding to the element in the ith column is obtained through a first function unit; the first function unit is configured to: obtaining a residual error of the ith column of the input matrix regressing the columns of the input matrix except the ith column according to the ith column of the input matrix, the columns of the input matrix except the ith column and the weight matrix; obtaining a weighting matrix according to the residual error, the tool variable matrix and the number of the user accounts of the plurality of user accounts; and obtaining an optimization function of the variable corresponding to the element of the ith column according to the residual error, the tool variable matrix and the weighting matrix.
Optionally, for a variable corresponding to an element in an ith column of the input matrix, when the variable corresponding to the element in the ith column is an exogenous variable, an optimization function of the variable corresponding to the element in the ith column is obtained through a second function unit; the second function unit is configured to: obtaining a residual error of the ith column of the input matrix regressing the columns of the input matrix except the ith column according to the ith column of the input matrix, the columns of the input matrix except the ith column and the weight matrix; and obtaining an optimization function of the variable corresponding to the element of the ith row according to the residual error.
According to a third aspect of the embodiments of the present disclosure, there is provided an electronic apparatus including: at least one processor; at least one memory storing computer-executable instructions, wherein the computer-executable instructions, when executed by the at least one processor, cause the at least one processor to perform an information processing method according to the present disclosure.
According to a fourth aspect of embodiments of the present disclosure, there is provided a computer-readable storage medium having stored therein instructions which, when executed by at least one processor, cause the at least one processor to execute an information processing method according to the present disclosure.
According to a fifth aspect of embodiments of the present disclosure, there is provided a computer program product comprising computer instructions which, when executed by at least one processor, implement an information processing method according to the present disclosure.
The technical scheme provided by the embodiment of the disclosure at least brings the following beneficial effects:
according to the information processing method and device disclosed by the invention, the causal relationship among a plurality of variables is obtained according to the experimental data set and the observed characteristic data set, and the experimental data can be introduced into the causal analysis, so that the problem of inaccurate inference of the causal association relationship among the operation information of the user account caused by missing data in the related art is solved, and the more accurate causal relationship among the user operation information is obtained.
In addition, according to the information processing method and the information processing device disclosed by the invention, the estimation error of a large weight matrix caused by the problem of missing variables in the related art can be corrected, and a stable causal relationship among the variables can be obtained.
In addition, according to the information processing method and apparatus of the present disclosure, a tool variable matrix can be obtained from an experimental data set, and applied to a predefined objective function, and an estimation bias of a weight matrix can be corrected by an uncorrelated condition of moments between an experiment and a residual using an external randomness of an experimental group.
It is to be understood that both the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the disclosure.
Drawings
The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present disclosure and, together with the description, serve to explain the principles of the disclosure and are not to be construed as limiting the disclosure.
Fig. 1 is a flow chart illustrating an information processing method according to an example embodiment.
Fig. 2 is an overall flowchart illustrating an information processing method according to an exemplary embodiment.
Fig. 3 is a block diagram illustrating an information processing apparatus according to an example embodiment.
Fig. 4 is a block diagram of an electronic device 400 according to an example embodiment.
Detailed Description
In order to make the technical solutions of the present disclosure better understood by those of ordinary skill in the art, the technical solutions in the embodiments of the present disclosure will be clearly and completely described below with reference to the accompanying drawings.
It should be noted that the terms "first," "second," and the like in the description and claims of the present disclosure and in the above-described drawings are used for distinguishing between similar elements and not necessarily for describing a particular sequential or chronological order. It is to be understood that the data so used is interchangeable under appropriate circumstances such that the embodiments of the disclosure described herein are capable of operation in sequences other than those illustrated or otherwise described herein. The embodiments described in the following examples do not represent all embodiments consistent with the present disclosure. Rather, they are merely examples of apparatus and methods consistent with certain aspects of the present disclosure, as detailed in the appended claims.
In this case, the expression "at least one of the items" in the present disclosure means a case where three types of parallel expressions "any one of the items", "a combination of any plural ones of the items", and "the entirety of the items" are included. For example, "include at least one of a and B" includes the following three cases in parallel: (1) comprises A; (2) comprises B; (3) including a and B. For another example, "at least one of the first step and the second step is performed", which means that the following three cases are juxtaposed: (1) executing the step one; (2) executing the step two; (3) and executing the step one and the step two.
For estimating the causal relationship between variables, methods such as regression are generally used, but these methods can only obtain the relationship of the direct influence of independent variables on dependent variables. For causal analysis with more independent variables, it is necessary to know whether there is a relationship between the independent variables and whether a variable has an indirect or direct influence on the dependent variable. To address this problem, the conventional method is based on searching all possible cause-and-effect graphs based on all nodes and possible relationships between the nodes, and generating a final cause-and-effect graph according to a certain criterion, where the nodes may be variables. However, this is a problem related to discrete search, which is a typical NP-hard problem, and the generation of the causal graph is very time-consuming according to the conventional method, and the existing computing resources are difficult to meet the operation requirement.
In the related art, a method for optimizing a Non-combinatorial search space for Structure learning by using a Trace number of a matrix and an Augmented Lagrangian function (Non-combinatorial Optimization Trace explicit and Augmented Lagrangian for Structure learning, notars) is an algorithm for learning a causal network relationship among multiple variables. The NOTAARS can convert the search interval of the discrete causal graph into a continuous search interval, namely, convert the problem of the discrete search into the problem of the continuous search, so that the generated causal graph meets the condition of directed no closed loop.
However, the NOTAARS approach is extremely hypothetical, given that the data for all variables related to the causal relationships in the causal graph must be observable and controllable. However, for some cases, data of all variables cannot be observed, and the data may affect the cause and effect relationship, in this case, the generation of the cause and effect graph by using the notaars may cause a large estimation error of the weight matrix due to the problem of missing variables, and the weight matrix characterizes the cause and effect relationship between the variables, which may make the cause and effect relationship inference in the cause and effect graph unstable, so that the generated cause and effect graph structure is unstable. For example, for a network platform, data of a part of variables such as daily idle time length of a user account cannot be observed, and in this case, when a user wishes to watch a video of the network platform in idle time, the recommended video duration of the network platform may be too long or too short due to the inability to observe the data of the daily idle time length of the user account, and the user's requirement may not be met. In order to better meet the requirements of users, inaccurate inference of causal association between operation information of user accounts caused by missing data needs to be solved when causal analysis of data related to a network platform is performed.
In order to solve the problems in the related art, the present disclosure provides an information processing method and apparatus, which obtain causal relationships among a plurality of variables according to an experimental data set and an observed characteristic data set, and can introduce experimental data into causal analysis, so that the problem of inaccurate inference of causal associations among operation information of user accounts due to missing data in the related art is solved, and a more accurate causal relationship among the user operation information is obtained.
Hereinafter, an information processing method and apparatus according to the present disclosure will be described in detail with reference to fig. 1 to 4.
Fig. 1 is a flow chart illustrating an information processing method according to an example embodiment. Referring to fig. 1, in step 101, an experimental data set and an observed characteristic data set may be obtained, where the characteristic data set includes characteristic data generated by each of a plurality of user accounts for each of a plurality of variables, the experimental data set is characteristic data obtained by performing experiments on the plurality of user accounts and experimental variables, the experimental variables include at least the plurality of variables, and the characteristic data is operation information of the user accounts on a predetermined platform.
According to an exemplary embodiment of the present disclosure, the experiment may be, but is not limited to, an AB experiment.
According to an example embodiment of the present disclosure, the variable may represent a category of feature data, for example, the variable may be a time period for viewing a video, and the feature data corresponding to the variable may include 30 minutes, 35 minutes, and the like.
At step 102, causal relationships among the plurality of variables may be obtained according to the experimental data set and the observed characteristic data set, wherein the causal relationships among the plurality of variables characterize correlations between different operation information.
According to an exemplary embodiment of the present disclosure, different operation information may be feature data generated by different variables. The different variable may be each of the plurality of variables, and the different operation information may be different operation information of the same user account.
According to an exemplary embodiment of the present disclosure, an input matrix may first be derived from an observed feature data set, and a tool variable matrix may be derived from an experimental data set. A weight matrix may then be derived from the input matrix and the tool variable matrix, wherein the weight matrix characterizes causal relationships between the plurality of variables. Here, the elements of each row of the input matrix are feature data generated by the same user account for each of a plurality of variables, the elements of each column of the input matrix are feature data generated by each of a plurality of user accounts for the same variable, the elements of each row of the tool variable matrix are feature data generated by the same user account for each of a plurality of experiments, the elements of each column of the tool variable matrix are feature data generated by each of a plurality of user accounts for the same experiment, and each of the plurality of experiments is performed for at least one of the experiment variables.
Here, each row of the input matrix may represent the characteristic data of one user account, the user accounts between the rows are not repeated, each column of the input matrix may represent the characteristic data of one variable, and the variables between the columns are not repeated. Each row of the tool variable matrix may represent characteristic data of one user account, user accounts between rows are not duplicated, each column of the tool variable matrix may represent characteristic data of one experiment, and experiments between columns are not duplicated.
According to an exemplary embodiment of the present disclosure, the predefined objective function may be optimized according to the input matrix and the tool variable matrix, resulting in a weight matrix that minimizes the value of the predefined objective function.
According to an exemplary embodiment of the present disclosure, before optimizing a predefined objective function according to an input matrix and a tool variable matrix to obtain a weight matrix that minimizes a value of the predefined objective function, values of elements in the weight matrix may be further limited according to a relationship between a plurality of predefined variables. For example, the plurality of variables may include the number of times a certain video is displayed and the number of times the certain video is viewed, the predefined relationship between the plurality of variables may include that the number of times the certain video is displayed determines the number of times the certain video is viewed, and the values of the elements related to the number of times the certain video is displayed and the number of times the certain video is viewed in the weight matrix may be limited based on the relationship. Here, the value of the element is limited, but not limited to, limiting the value range of the element, and setting the element to a preset value.
According to the exemplary embodiment of the present disclosure, after the values of the elements in the weight matrix are limited according to the predefined relationship among the plurality of variables, the predefined objective function may be optimized to obtain the weight matrix that minimizes the value of the predefined objective function, and the values of the elements in the obtained weight matrix still conform to the limited range. For example, the first element in the weight matrix is set to a preset value according to a predefined relationship among a plurality of variables, and the first element in the weight matrix that minimizes the value of the predefined objective function, which is finally obtained, is still the preset value.
Here, the predefined objective function may include an optimization term and a penalty term.
For example, the predefined objective function may be represented by equation (1):
F(W)＝l(W；X)optimization term+λ||W||1 (1)
Wherein F (W) is a predefined objective function, l (W; X)Optimization termTo optimize the term, lambda | W | ceiling1For penalty term, W is a weighting matrix of k × k dimension, X is an input matrix of N × k dimension, N, k are positive integers, N is the number of user accounts of a plurality of user accounts, k is the number of variables of a plurality of variables, λ is an adjustable parameter, | | W | | Y1An L-1penalty term (L-1penalty) for W. Note that L-1penalty terms can be used to penalize sparse matrices, and λ can be determined by cross-validation (cross-validation).
It should be noted that, before the predefined objective function is optimized, each element in the weight matrix exists in the form of an initial value, and the initial value of each element in the weight matrix may be a preset custom value.
For the optimization term in the predefined objective function, NOTAARS in the related art can be expressed by the following formula (2):
wherein l (W; X)Optimization termFor optimization, W is a k × k dimensional weight matrix, X is an N × k dimensional input matrix, N, k are positive integers, N is the number of user accounts of a plurality of user accounts, k is the number of variables of a plurality of variables, | X-XW | | Y cellsFThe Frobenius norm (frobenius norm) represents X-XW, which is the product of X and W in the N X k dimensions.
It can be seen that the optimization term represented in equation (2) does not classify variables corresponding to elements of each column of the input matrix, and based on this, in an exemplary embodiment of the present disclosure, the optimization term may be a sum of an optimization function of each endogenous variable and an optimization function of each exogenous variable, wherein the endogenous variable and the exogenous variable may be variables of a plurality of variables, the endogenous variable is related to an unobserved variable, the exogenous variable is not related to an unobserved variable, and the unobserved variable affects a causal relationship between the plurality of variables and is not a variable of the plurality of variables. For example, an unobserved variable may be a variable that affects causal relationships between multiple variables, but for which corresponding feature data cannot be observed.
According to an exemplary embodiment of the present disclosure, for a variable corresponding to an element in an ith column of an input matrix, in a case where the variable corresponding to the element in the ith column is an endogenous variable, an optimization function of the variable corresponding to the element in the ith column is obtained by: firstly, the residual error of the ith column of the input matrix regressing the columns of the input matrix except the ith column can be obtained according to the ith column of the input matrix, the columns of the input matrix except the ith column and the weight matrix. And then obtaining a weighting matrix according to the residual error, the tool variable matrix and the number of the user accounts of the plurality of user accounts. And finally, obtaining an optimization function of the variable corresponding to the element of the ith column according to the residual error, the tool variable matrix and the weighting matrix.
For example, for a variable corresponding to an element in the ith column of the input matrix, when the variable corresponding to the element in the ith column is an endogenous variable, the optimization function of the variable corresponding to the element in the ith column is expressed by the following formula (3):
wherein the content of the first and second substances,
ui＝Xi-∑-ix-iWi，-ii is more than or equal to 1 and less than or equal to k, L (W; X) is an optimization function of variables corresponding to elements in the ith column, W is a k multiplied by k dimensional weight matrix, X is an N multiplied by k dimensional input matrix, Z is an N multiplied by L dimensional tool variable matrix, N, k and L are positive integers, N is the number of user accounts of a plurality of user accounts, k is the number of variables of a plurality of variables, L is the number of experiments of a plurality of experiments, L is more than or equal to k, Z' is a transposed matrix of Z, X is equal to or greater than the number of experiments, andiis the ith column of X, X-iIs X except the ith column, Wi，-iIs the ith row of W and the columns of W other than the ith column, MkIs a weighting matrix, uiIs the residual of the ith column of X regressing the columns of X except the ith column, ui' is the transposed matrix of ui.
According to an exemplary embodiment of the present disclosure, for a variable corresponding to an element in an ith column of an input matrix, in the case that the variable corresponding to the element in the ith column is an exogenous variable, an optimization function of the variable corresponding to the element in the ith column is obtained by: obtaining a residual error of the ith column of the input matrix regressing the columns of the input matrix except the ith column according to the ith column of the input matrix, the columns of the input matrix except the ith column and the weight matrix; and obtaining an optimization function of the variable corresponding to the element in the ith column according to the residual error.
For example, for a variable corresponding to an element in the ith column of the input matrix, in the case where the variable corresponding to the element in the ith column is an exogenous variable, the optimization function of the variable corresponding to the element in the ith column is expressed by the following formula (4):
wherein i is more than or equal to 1 and less than or equal to k, l (W; X) is an optimization function of variables corresponding to elements in the ith column, W is a k multiplied by k dimensional weight matrix, X is an N multiplied by k dimensional input matrix, N, k are positive integers, N is the number of user accounts of a plurality of user accounts, k is the number of variables of a plurality of variables, and X is the number of variables of a plurality of variablesiIs the ith column of X, X-iIs X except the ith column, Wi，-iIs the ith row of W and the columns of W other than the ith column. In addition, (X)i-∑-iX-iWi，-i) The ith column for X regresses the residuals of the columns of X except the ith column.
According to an example embodiment of the present disclosure, a causal graph may be generated for a plurality of variables based on a plurality of variables and causal relationships between the plurality of variables.
Fig. 2 is an overall flowchart illustrating an information processing method according to an exemplary embodiment. An overall description is made below of an information processing method in an exemplary embodiment of the present disclosure with reference to fig. 2.
First, an experimental data set and an observed characteristic data set may be obtained. The feature data set includes feature data generated by each of the plurality of user accounts for each of the plurality of variables, the experimental data set is feature data obtained by performing experiments on the plurality of user accounts and experimental variables, the experimental variables at least include the plurality of variables, and the feature data may be operation information of the user accounts on a predetermined platform.
Here, an input matrix may be derived from the observed feature data set, and a tool variable matrix may be derived from the experimental data set.
Secondly, the values of the elements in the weight matrix contained in the predefined objective function can be limited according to the predefined relationship among a plurality of variables.
Then, the predefined objective function may be optimized according to the input matrix and the tool variable matrix to obtain a weight matrix that minimizes a value of the predefined objective function, wherein the weight matrix characterizes causal relationships between the plurality of variables. The causal relationship between multiple variables reflects the correlation between different operational information.
Here, the predefined objective function is optimized, and the above equations (1), (3) and (4) may be referred to.
Finally, a causal graph may be generated for the plurality of variables based on the plurality of variables and causal relationships between the plurality of variables.
In the following, an information processing method according to the present disclosure is explained as an exemplary example.
First, an experimental data set and an observed characteristic data set may be obtained. The feature data set comprises feature data of each user account in a plurality of user accounts aiming at a plurality of variables such as average time length of watching videos on a preset platform every day and average frequency of clicking videos displayed on a homepage on the preset platform every day, the experiment data set is feature data obtained by carrying out experiments on a plurality of user accounts and experiment variables, and the experiment variables can be average time length of watching videos on the preset platform every day, average time length of watching live broadcasts on the preset platform every day, average frequency of clicking videos displayed on the homepage on the preset platform every day and average frequency of clicking live broadcasts displayed on the preset platform every day.
It should be noted that the feature data may be operation information of the user account on a predetermined platform, for example, the feature data may be that the average time duration of watching a video on the predetermined platform per day of the first user account is 30 minutes, the average time duration of watching a video on the predetermined platform per day of the second user account is 50 minutes, the average number of times of clicking a video displayed on the homepage on the predetermined platform per day of the first user account is 10 times, the average number of times of clicking a video displayed on the homepage on the predetermined platform per day of the second user account is 20 times, and the like.
Here, an input matrix may be derived from the observed feature data set, and a tool variable matrix may be derived from the experimental data set.
Secondly, the values of the elements in the weight matrix contained in the predefined objective function can be limited according to the predefined relationship among a plurality of variables. For example, the predefined relationship between the variables may be that the average number of times that the video displayed on the home page is clicked on the predetermined platform per day and the average duration of watching the video on the predetermined platform per day are in a positive correlation, and then, values of elements related to the above-described relationship in the weight matrix may be limited, for example, the elements related to the above-described relationship in the weight matrix are elements in the a-th row and the b-th column, and the elements in the a-th row and the b-th column of the weight matrix may be set to preset values.
Finally, the predefined objective function may be optimized according to the input matrix and the tool variable matrix to obtain a weight matrix that minimizes a value of the predefined objective function, wherein the weight matrix characterizes causal relationships among the plurality of variables.
It should be noted that the causal relationship between the variables reflects the correlation between the average time length of watching the video on the predetermined platform per day and the average number of times of clicking the video displayed on the homepage on the predetermined platform per day for each user account.
For example, a causal relationship between the plurality of variables may be that the average number of times per user account a displayed video clicks a home page on a predetermined platform per day is the reason, and the average length of time a video is viewed on a predetermined platform per day is the result.
Here, the predefined objective function is optimized, and the above equations (1), (3) and (4) may be referred to.
It should be further noted that the exemplary embodiments of the present disclosure are only examples of determining a causal relationship between an average time length of each user account viewing a video on a predetermined platform every day and an average number of times of clicking a home page displayed on the predetermined platform every day, and the present disclosure also protects a causal relationship between other operation information of the user accounts on the predetermined platform, such as the number of times of viewing the video on the predetermined platform and the number of videos collected on the predetermined platform, and will not be described herein again.
Fig. 3 is a block diagram illustrating an information processing apparatus according to an example embodiment. Referring to fig. 3, the information processing apparatus 300 includes an acquisition unit 301 and a determination unit 302.
The obtaining unit 301 may obtain an experimental data set and an observed feature data set, where the feature data set includes feature data generated by each of a plurality of user accounts for each of a plurality of variables, the experimental data set is feature data obtained by performing experiments on the plurality of user accounts and the experimental variables, the experimental variables at least include the plurality of variables, and the feature data is operation information of the user accounts on a predetermined platform.
According to an exemplary embodiment of the present disclosure, the experiment may be, but is not limited to, an AB experiment.
According to an example embodiment of the present disclosure, the variable may represent a category of feature data, for example, the variable may be a time period for viewing a video, and the feature data corresponding to the variable may include 30 minutes, 35 minutes, and the like.
The determining unit 302 may obtain a causal relationship between multiple variables according to the experimental data set and the observed characteristic data set, where the causal relationship between the multiple variables represents an association relationship between different operation information.
According to an exemplary embodiment of the present disclosure, different operation information may be feature data generated by different variables. The different variable may be each of the plurality of variables, and the different operation information may be different operation information of the same user account.
According to an exemplary embodiment of the present disclosure, the determining unit 302 may first derive an input matrix from the observed feature data set and a tool variable matrix from the experimental data set. The determining unit 302 may then derive a weight matrix from the input matrix and the tool variable matrix, wherein the weight matrix characterizes causal relationships between the plurality of variables. Here, the elements of each row of the input matrix are feature data generated by the same user account for each of a plurality of variables, the elements of each column of the input matrix are feature data generated by each of a plurality of user accounts for the same variable, the elements of each row of the tool variable matrix are feature data generated by the same user account for each of a plurality of experiments, the elements of each column of the tool variable matrix are feature data generated by each of a plurality of user accounts for the same experiment, and each of the plurality of experiments is performed for at least one of the experiment variables.
Here, each row of the input matrix may represent the characteristic data of one user account, the user accounts between the rows are not repeated, each column of the input matrix may represent the characteristic data of one variable, and the variables between the columns are not repeated. Each row of the tool variable matrix may represent characteristic data of one user account, user accounts between rows are not duplicated, each column of the tool variable matrix may represent characteristic data of one experiment, and experiments between columns are not duplicated.
According to an exemplary embodiment of the present disclosure, the determining unit 302 may optimize the predefined objective function according to the input matrix and the tool variable matrix, to obtain a weight matrix that minimizes a value of the predefined objective function.
According to an exemplary embodiment of the present disclosure, the method further includes an adjusting unit, where the adjusting unit may limit values of elements in the weight matrix according to a relationship between a plurality of predefined variables before optimizing the predefined objective function according to the input matrix and the tool variable matrix to obtain the weight matrix that minimizes a value of the predefined objective function. For example, the plurality of variables may include the number of times a certain video is displayed and the number of times the certain video is viewed, the predefined relationship between the plurality of variables may include that the number of times the certain video is displayed determines the number of times the certain video is viewed, and the adjusting unit may limit the value of an element related to the number of times the certain video is displayed and the number of times the certain video is viewed in the weight matrix based on the relationship. Here, the value of the element is limited, but not limited to, limiting the value range of the element, and setting the element to a preset value.
According to an exemplary embodiment of the present disclosure, after the adjusting unit limits the values of the elements in the weight matrix according to the predefined relationship among the multiple variables, the determining unit 302 may optimize the predefined objective function to obtain the weight matrix that minimizes the value of the predefined objective function, and the values of the elements in the obtained weight matrix still conform to the limited range. For example, the first element in the weight matrix is set to a preset value according to a predefined relationship among a plurality of variables, and the first element in the weight matrix that minimizes the value of the predefined objective function, which is finally obtained, is still the preset value.
Here, the predefined objective function may include an optimization term and a penalty term.
For example, the predefined objective function may be represented as equation (1) above.
It should be noted that, before the predefined objective function is optimized, each element in the weight matrix exists in the form of an initial value, and the initial value of each element in the weight matrix may be a preset custom value.
For the optimization term in the predefined objective function, NOTAARS in the related art can be expressed by the above equation (2).
The optimization term represented in equation (2) does not classify variables corresponding to elements of each column of the input matrix, and based on this, in an exemplary embodiment of the present disclosure, the optimization term may be a sum of an optimization function of each endogenous variable and an optimization function of each exogenous variable, wherein the endogenous variable and the exogenous variable may be variables of a plurality of variables, the endogenous variable is related to an unobserved variable, the exogenous variable is unrelated to an unobserved variable, and the unobserved variable affects a causal relationship between the plurality of variables and is not a variable of the plurality of variables. For example, an unobserved variable may be a variable that affects causal relationships between multiple variables, but for which corresponding feature data cannot be observed.
According to an exemplary embodiment of the present disclosure, for a variable corresponding to an element in an ith column of an input matrix, when the variable corresponding to the element in the ith column is an endogenous variable, an optimization function of the variable corresponding to the element in the ith column is obtained by a first function unit; the first function unit is configured to: obtaining a residual error of the ith column of the input matrix regressing the columns of the input matrix except the ith column according to the ith column of the input matrix, the columns of the input matrix except the ith column and the weight matrix; obtaining a weighting matrix according to the residual error, the tool variable matrix and the number of the user accounts of the plurality of user accounts; and obtaining an optimization function of the variable corresponding to the element of the ith column according to the residual error, the tool variable matrix and the weighting matrix.
For example, in the case where the variable corresponding to the element in the ith column of the input matrix is an endogenous variable, the optimization function of the variable corresponding to the element in the ith column is expressed by the above expression (3).
According to an exemplary embodiment of the present disclosure, for a variable corresponding to an element in an ith column of an input matrix, when the variable corresponding to the element in the ith column is an exogenous variable, an optimization function of the variable corresponding to the element in the ith column is obtained by a second function unit; the second function unit is configured to: obtaining a residual error of the ith column of the input matrix regressing the columns of the input matrix except the ith column according to the ith column of the input matrix, the columns of the input matrix except the ith column and the weight matrix; and obtaining an optimization function of the variable corresponding to the element in the ith column according to the residual error.
For example, in the case where the variable corresponding to the element in the ith column of the input matrix is an exogenous variable, the optimization function of the variable corresponding to the element in the ith column is expressed by the above expression (4).
According to an exemplary embodiment of the present disclosure, the information processing may further include a generation unit that may generate a causal graph with respect to the plurality of variables according to the plurality of variables and causal relationships between the plurality of variables.
Fig. 4 is a block diagram of an electronic device 400 according to an example embodiment.
Referring to fig. 4, the electronic device 400 includes at least one memory 401 and at least one processor 402, the at least one memory 401 having stored therein a set of computer-executable instructions that, when executed by the at least one processor 402, perform an information processing method according to an exemplary embodiment of the present disclosure.
By way of example, the electronic device 400 may be a PC computer, tablet device, personal digital assistant, smartphone, or other device capable of executing the set of instructions described above. Here, the electronic device 400 need not be a single electronic device, but can be any collection of devices or circuits that can execute the above instructions (or sets of instructions) individually or in combination. The electronic device 400 may also be part of an integrated control system or system manager, or may be configured as a portable electronic device that interfaces with local or remote (e.g., via wireless transmission).
In the electronic device 400, the processor 402 may include a Central Processing Unit (CPU), a Graphics Processing Unit (GPU), a programmable logic device, a special purpose processor system, a microcontroller, or a microprocessor. By way of example, and not limitation, processors may also include analog processors, digital processors, microprocessors, multi-core processors, processor arrays, network processors, and the like.
The processor 402 may execute instructions or code stored in the memory 401, wherein the memory 401 may also store data. The instructions and data may also be transmitted or received over a network via a network interface device, which may employ any known transmission protocol.
The memory 401 may be integrated with the processor 402, for example, by having RAM or flash memory disposed within an integrated circuit microprocessor or the like. Further, memory 401 may comprise a stand-alone device, such as an external disk drive, storage array, or any other storage device usable by a database system. The memory 401 and the processor 402 may be operatively coupled or may communicate with each other, such as through I/O ports, network connections, etc., so that the processor 402 can read files stored in the memory.
In addition, the electronic device 400 may also include a video display (such as a liquid crystal display) and a user interaction interface (such as a keyboard, mouse, touch input device, etc.). All components of electronic device 400 may be connected to each other via a bus and/or a network.
According to an exemplary embodiment of the present disclosure, there may also be provided a computer-readable storage medium, wherein when instructions stored in the computer-readable storage medium are executed by at least one processor, the at least one processor is caused to perform an information processing method according to an exemplary embodiment of the present disclosure. Examples of the computer-readable storage medium herein include: read-only memory (ROM), random-access programmable read-only memory (PROM), electrically erasable programmable read-only memory (EEPROM), random-access memory (RAM), dynamic random-access memory (DRAM), static random-access memory (SRAM), flash memory, non-volatile memory, CD-ROM, CD-R, CD + R, CD-RW, CD + RW, DVD-ROM, DVD-R, DVD + R, DVD-RW, DVD + RW, DVD-RAM, BD-ROM, BD-R, BD-R LTH, BD-RE, Blu-ray or compact disc memory, Hard Disk Drive (HDD), solid-state drive (SSD), card-type memory (such as a multimedia card, a Secure Digital (SD) card or a extreme digital (XD) card), magnetic tape, a floppy disk, a magneto-optical data storage device, an optical data storage device, a hard disk, a magnetic tape, a magneto-optical data storage device, a hard disk, a magnetic tape, a magnetic data storage device, a magnetic tape, a magnetic data storage device, a magnetic tape, a magnetic data storage device, a magnetic tape, a magnetic data storage device, a magnetic tape, a magnetic data storage device, A solid state disk, and any other device configured to store and provide a computer program and any associated data, data files, and data structures to a processor or computer in a non-transitory manner such that the processor or computer can execute the computer program. The computer program in the computer-readable storage medium described above can be run in an environment deployed in a computer apparatus, such as a client, a host, a proxy device, a server, and the like, and further, in one example, the computer program and any associated data, data files, and data structures are distributed across a networked computer system such that the computer program and any associated data, data files, and data structures are stored, accessed, and executed in a distributed fashion by one or more processors or computers.
According to an exemplary embodiment of the present disclosure, there may also be provided a computer program product in which instructions are executable by a processor of a computer apparatus to perform an information processing method according to an exemplary embodiment of the present disclosure.
According to the information processing method and device disclosed by the invention, the causal relationship among a plurality of variables is obtained according to the experimental data set and the observed characteristic data set, and the experimental data can be introduced into the causal analysis, so that the problem of inaccurate inference of the causal association relationship among the operation information of the user account caused by missing data in the related art is solved, and the more accurate causal relationship among the user operation information is obtained.
In addition, according to the information processing method and the information processing device disclosed by the invention, the estimation error of a large weight matrix caused by the problem of missing variables in the related art can be corrected, and a stable causal relationship among the variables can be obtained.
In addition, according to the information processing method and apparatus of the present disclosure, a tool variable matrix can be obtained from an experimental data set, and applied to a predefined objective function, and an estimation bias of a weight matrix can be corrected by an uncorrelated condition of moments between an experiment and a residual using an external randomness of an experimental group.
Other embodiments of the disclosure will be apparent to those skilled in the art from consideration of the specification and practice of the disclosure disclosed herein. This application is intended to cover any variations, uses, or adaptations of the disclosure following, in general, the principles of the disclosure and including such departures from the present disclosure as come within known or customary practice within the art to which the disclosure pertains. It is intended that the specification and examples be considered as exemplary only, with a true scope and spirit of the disclosure being indicated by the following claims.
It will be understood that the present disclosure is not limited to the precise arrangements described above and shown in the drawings and that various modifications and changes may be made without departing from the scope thereof. The scope of the present disclosure is limited only by the appended claims.
Claims (10)
1. An information processing method characterized by comprising:
acquiring an experimental data set and an observed characteristic data set, wherein the characteristic data set comprises characteristic data generated by each user account in a plurality of user accounts for each variable in a plurality of variables, the experimental data set is characteristic data obtained by performing experiments on the plurality of user accounts and experimental variables, the experimental variables at least comprise the plurality of variables, and the characteristic data is operation information of the user accounts on a predetermined platform;
and obtaining the causal relationship among the variables according to the experimental data set and the observed characteristic data set, wherein the causal relationship among the variables represents the association relationship among different operation information.
2. The information processing method of claim 1, wherein said deriving causal relationships between said plurality of variables from said experimental data set and said observed characteristic data set comprises:
deriving an input matrix from said observed feature data set, and deriving a tool variable matrix from said experimental data set,
wherein elements of each row of the input matrix are feature data generated by the same user account for each of the plurality of variables, elements of each column of the input matrix are feature data generated by each of the plurality of user accounts for the same variable, elements of each row of the tool variable matrix are feature data generated by the same user account for each of a plurality of experiments, elements of each column of the tool variable matrix are feature data generated by each of the plurality of user accounts for the same experiment, and each of the plurality of experiments is conducted for at least one of the experimental variables;
and obtaining a weight matrix according to the input matrix and the tool variable matrix, wherein the weight matrix represents the causal relationship among the variables.
3. The information processing method of claim 2, wherein said deriving a weight matrix from said input matrix and said tool variable matrix comprises:
and optimizing a predefined objective function according to the input matrix and the tool variable matrix to obtain a weight matrix which enables the value of the predefined objective function to be minimum.
4. The information processing method of claim 3, wherein before the optimizing a predefined objective function based on the input matrix and the tool variable matrix to obtain a weight matrix that minimizes a value of the predefined objective function, further comprises:
and limiting the values of the elements in the weight matrix according to the predefined relationship among the variables.
5. The information processing method according to claim 3, wherein the predefined objective function includes an optimization term that is a sum value of an optimization function of each endogenous variable and an optimization function of each exogenous variable,
wherein the endogenous variable and the exogenous variable are variables of the plurality of variables, the endogenous variable is related to an unobserved variable, the exogenous variable is unrelated to the unobserved variable, and the unobserved variable affects a causal relationship between the plurality of variables and is not a variable of the plurality of variables.
6. The information processing method according to claim 5, wherein, for the variable corresponding to the element in the ith column of the input matrix, in a case where the variable corresponding to the element in the ith column is an endogenous variable, the optimization function of the variable corresponding to the element in the ith column is obtained by:
obtaining a residual error of the ith column of the input matrix regressing the columns of the input matrix except the ith column according to the ith column of the input matrix, the columns of the input matrix except the ith column and the weight matrix;
obtaining a weighting matrix according to the residual error, the tool variable matrix and the number of the user accounts of the plurality of user accounts;
and obtaining an optimization function of the variable corresponding to the element of the ith column according to the residual error, the tool variable matrix and the weighting matrix.
7. An information processing apparatus characterized by comprising:
an acquisition unit configured to: acquiring an experimental data set and an observed characteristic data set, wherein the characteristic data set comprises characteristic data generated by each user account in a plurality of user accounts for each variable in a plurality of variables, the experimental data set is characteristic data obtained by performing experiments on the plurality of user accounts and experimental variables, the experimental variables at least comprise the plurality of variables, and the characteristic data is operation information of the user accounts on a predetermined platform;
a determination unit configured to: and obtaining the causal relationship among the variables according to the experimental data set and the observed characteristic data set, wherein the causal relationship among the variables represents the association relationship among different operation information.
8. An electronic device, comprising:
at least one processor;
at least one memory storing computer-executable instructions,
wherein the computer-executable instructions, when executed by the at least one processor, cause the at least one processor to perform the information processing method of any one of claims 1 to 6.
9. A computer-readable storage medium, wherein instructions stored in the computer-readable storage medium, when executed by at least one processor, cause the at least one processor to perform the information processing method of any one of claims 1 to 6.
10. A computer program product comprising computer instructions, characterized in that the computer instructions, when executed by at least one processor, implement the information processing method of any of claims 1 to 6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111510017.8A CN114186096A (en) | 2021-12-10 | 2021-12-10 | Information processing method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111510017.8A CN114186096A (en) | 2021-12-10 | 2021-12-10 | Information processing method and device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114186096A true CN114186096A (en) | 2022-03-15 |
Family
ID=80543224
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111510017.8A Pending CN114186096A (en) | 2021-12-10 | 2021-12-10 | Information processing method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114186096A (en) |
Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20140289174A1 (en) * | 2013-03-15 | 2014-09-25 | Alexander Statnikov | Data Analysis Computer System and Method For Causal Discovery with Experimentation Optimization |
| US20160292248A1 (en) * | 2013-11-22 | 2016-10-06 | Cambridge Social Science Decision Lab Inc. | Methods, systems, and articles of manufacture for the management and identification of causal knowledge |
| US20170124240A1 (en) * | 2015-10-28 | 2017-05-04 | International Business Machines Corporation | Scalable and automated identification of unobservability causality in logic optimization flows |
| CN109598346A (en) * | 2017-09-30 | 2019-04-09 | 日本电气株式会社 | For estimating the causal methods, devices and systems between observational variable |
| US20200279417A1 (en) * | 2019-02-28 | 2020-09-03 | Babylon Partners Limited | Generating a probabilistic graphical model with causal information |
| CN112380273A (en) * | 2020-11-11 | 2021-02-19 | 北京达佳互联信息技术有限公司 | Data estimation method and device |
| CN112633503A (en) * | 2020-12-16 | 2021-04-09 | 浙江大学 | Tool variable generation and counterfactual reasoning method and device based on neural network |
| CN112685674A (en) * | 2020-12-30 | 2021-04-20 | 百果园技术(新加坡)有限公司 | Feature evaluation method and device influencing user retention |
| CN113256353A (en) * | 2021-07-05 | 2021-08-13 | 北京达佳互联信息技术有限公司 | Business data processing method and construction method of business data processing system |
| US20210271984A1 (en) * | 2020-02-28 | 2021-09-02 | Alibaba Group Holding Limited | Regression modeling of sparse acyclic graphs in time series causal inference |
| CN113409901A (en) * | 2021-06-29 | 2021-09-17 | 南华大学 | Causal inference method and system for cascade medical observation data |
-
2021
- 2021-12-10 CN CN202111510017.8A patent/CN114186096A/en active Pending
Patent Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20140289174A1 (en) * | 2013-03-15 | 2014-09-25 | Alexander Statnikov | Data Analysis Computer System and Method For Causal Discovery with Experimentation Optimization |
| US20160292248A1 (en) * | 2013-11-22 | 2016-10-06 | Cambridge Social Science Decision Lab Inc. | Methods, systems, and articles of manufacture for the management and identification of causal knowledge |
| US20170124240A1 (en) * | 2015-10-28 | 2017-05-04 | International Business Machines Corporation | Scalable and automated identification of unobservability causality in logic optimization flows |
| CN109598346A (en) * | 2017-09-30 | 2019-04-09 | 日本电气株式会社 | For estimating the causal methods, devices and systems between observational variable |
| US20200279417A1 (en) * | 2019-02-28 | 2020-09-03 | Babylon Partners Limited | Generating a probabilistic graphical model with causal information |
| US20210271984A1 (en) * | 2020-02-28 | 2021-09-02 | Alibaba Group Holding Limited | Regression modeling of sparse acyclic graphs in time series causal inference |
| CN112380273A (en) * | 2020-11-11 | 2021-02-19 | 北京达佳互联信息技术有限公司 | Data estimation method and device |
| CN112633503A (en) * | 2020-12-16 | 2021-04-09 | 浙江大学 | Tool variable generation and counterfactual reasoning method and device based on neural network |
| CN112685674A (en) * | 2020-12-30 | 2021-04-20 | 百果园技术(新加坡)有限公司 | Feature evaluation method and device influencing user retention |
| CN113409901A (en) * | 2021-06-29 | 2021-09-17 | 南华大学 | Causal inference method and system for cascade medical observation data |
| CN113256353A (en) * | 2021-07-05 | 2021-08-13 | 北京达佳互联信息技术有限公司 | Business data processing method and construction method of business data processing system |
Non-Patent Citations (5)
| Title |
|---|
| P. GUO 等: "Parallel Gradient Boosting based Granger Causality Learning", 2019 IEEE INTERNATIONAL CONFERENCE ON BIG DATA (BIG DATA), 12 December 2019 (2019-12-12), pages 2845 - 2854, XP033721093, DOI: 10.1109/BigData47090.2019.9005690 * |
| 单娜 等: "带有测量误差协变量和不可观测反应变量下的因果推断", 吉林师范大学学报(自然科学版), vol. 41, no. 01, 4 February 2020 (2020-02-04), pages 62 - 66 * |
| 苗旺 等: "因果推断的统计方法", 中国科学:数学, vol. 48, no. 12, 13 December 2018 (2018-12-13), pages 1753 - 1778 * |
| 赵慧 等: "含隐变量和选择偏差的图模型中的因果推断", 北京大学学报(自然科学版), vol. 42, no. 05, 20 September 2006 (2006-09-20), pages 584 - 589 * |
| 陈薇: "面向不完全观察数据的混淆因子检测与因果结构学习研究", 中国博士学位论文全文数据库 基础科学辑, no. 02, 15 February 2021 (2021-02-15), pages 002 - 48 * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11327934B2 (en) | Systems and methods for cleansing automated robotic traffic from sets of usage logs | |
| Asparouhov et al. | Structural equation models and mixture models with continuous nonnormal skewed distributions | |
| WO2019129060A1 (en) | Method and system for automatically generating machine learning sample | |
| US9904584B2 (en) | Performance anomaly diagnosis | |
| Novi Inverardi et al. | Maximum entropy density estimation from fractional moments | |
| US20210136098A1 (en) | Root cause analysis in multivariate unsupervised anomaly detection | |
| Thomas | Making sense of snapshot data: ergodic principle for clonal cell populations | |
| US9071510B2 (en) | Determining root causes of network issues | |
| Daly et al. | Hodgkin–Huxley revisited: reparametrization and identifiability analysis of the classic action potential model with approximate Bayesian methods | |
| US10853373B1 (en) | Dynamic data layouts for time-series data | |
| WO2013067461A2 (en) | Identifying associations in data | |
| CN113822440A (en) | Method and system for determining feature importance of machine learning samples | |
| US20160255109A1 (en) | Detection method and apparatus | |
| CN111242310A (en) | Feature validity evaluation method and device, electronic equipment and storage medium | |
| Lorenzo-Seva | Horn’s Parallel Analysis for selecting the number of dimensions in Correspondence Analysis | |
| US20120283991A1 (en) | Method and System for Online Detection of Multi-Component Interactions in Computing Systems | |
| Khan et al. | How fluxional reactants limit the accuracy/efficiency of infrequent metadynamics | |
| US10509800B2 (en) | Visually interactive identification of a cohort of data objects similar to a query based on domain knowledge | |
| US8971644B1 (en) | System and method for determining an annotation for an image | |
| JP2012194741A (en) | Prediction device of missing value in matrix data, method for calculating missing value prediction, and missing value prediction program | |
| JP2018195099A (en) | Character analyzing device, device for generating model for character analysis, and character analyzing program | |
| US9922071B2 (en) | Isolation anomaly quantification through heuristical pattern detection | |
| CN114186096A (en) | Information processing method and device | |
| CN116467975A (en) | Data processing method, device, electronic equipment and storage medium | |
| JP6091448B2 (en) | Topic modeling device, topic modeling method, topic modeling program |
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 |