Missing data complementary device and missing data complementary method

Abstract

To provide a missing data complementary device and a missing data complementary method which improve the overall prediction accuracy of learning data including missing data, while reducing the missing data to be complemented more than before.SOLUTION: A missing data complementary device includes: a complementation priority flagging section for giving a complementation priority flag to at least one entry included in a predetermined ratio defined by a complementation amount adjustment parameter in each datum of a specific string data item; and a priority order determination section for counting the number of the complementation priority flags given to each entry in each entry included in a data table, so as to determine integrated complementation priority rank order in descending order of the number of the complementation priority flags.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to a missing data complementation device and a missing data complementation method, and is suitable for application to a missing data complementation device relating to a technique for complementing missing data, for example. [Background technology]

[0002] In recent years, advances in artificial intelligence technology have led to active research into machine learning using training data. Such training data may contain missing data, and various studies have been conducted to address such cases. Patent Document 1 discloses a technique for imputing missing data as a preprocessing step for machine learning. The technique disclosed in Patent Document 1 aims to improve the accuracy of imputing missing data. Specifically, a correlation matrix calculation unit first calculates a correlation matrix between attributes using all training records, and a regression imputation unit performs regression imputation using attributes whose absolute correlation values ​​are greater than a correlation threshold for the missing attributes, if any. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Publication No. 2020-154828 Summary of the Invention [Problem to be solved by the invention]

[0004] However, the technology disclosed in Patent Document 1 only performs regression imputation for missing data as described above, and has not considered how much missing data needs to be imputed to improve the overall prediction accuracy of training data including missing data.

[0005] The present invention has been made in consideration of the above points, and aims to propose a missing data imputation device and a missing data imputation method that can improve the overall prediction accuracy of training data including missing data while minimizing the amount of missing data to be imputed. [Means for solving the problem]

[0006] In order to solve the above problem, the present invention provides a data table having a plurality of column data items defined in the column direction and a plurality of entries in the row direction, each of which is composed of data for each of the plurality of column data items, wherein missing data exists in a specific column data item in some of the entries; a parameter definition unit defining a completion amount adjustment parameter used to adjust a proportion of data for the specific column data item that needs to be completed; a missing data prediction unit predicting the missing data for the specific column data item based on data for column data items other than the specific column data item in the plurality of entries that make up the data table and the data for the specific column data item; an extraction unit extracting at least one entry that falls within a predetermined proportion defined by the completion amount adjustment parameter from each piece of data for the specific column data item; a completion priority flag assignment unit assigning a completion priority flag to at least one entry that falls within the predetermined proportion; and a priority ranking determination unit that counts the number of entries in the data table to which the completion priority flag has been assigned, and determines an integrated completion priority ranking in descending order of the number of entries with the largest number of completion priority flags.

[0007] Further, in the present invention, a data table has a plurality of column data items defined in the column direction and a plurality of entries configured with each data of the plurality of column data items in the row direction, and missing data exists in a specific column data item in some of the entries, in which a parameter definition step defines in a parameter definition section a complement amount adjustment parameter used to adjust the proportion of data that needs to be complemented for the specific column data item; and a missing data prediction section calculates the proportion of the specific column data based on data of column data items other than the specific column data item and the data of the specific column data item in the plurality of entries that make up the data table. The data table method includes a missing data prediction step of predicting the missing data of an item, an extraction step in which an extraction unit extracts at least one of the entries included in a predetermined ratio defined by the completion amount adjustment parameter from each data of the specific column data item, a completion priority flag assignment step in which a completion priority flag assignment unit assigns a completion priority flag to at least one of the entries included in the predetermined ratio, and a priority ranking determination step in which a priority ranking determination unit counts the number of entries included in the data table to which the completion priority flag is assigned, and determines an integrated completion priority ranking in descending order of the number of completion priority flags. [Effects of the Invention]

[0008] According to the present invention, it is possible to improve the overall prediction accuracy of training data including missing data while minimizing the amount of missing data to be complemented. [Brief explanation of the drawings]

[0009] [Figure 1] 1 is a system configuration diagram showing an example of the configuration of a missing data compensation device according to a first embodiment. [Figure 2] 10 is a flowchart illustrating an example of a procedure for missing data filling processing. [Figure 3] 3 is a flowchart showing an example of a procedure for processing missing data shown in FIG. 2. [Figure 4]FIG. 10 is a diagram illustrating an example of the contents of a data table. [Figure 5] FIG. 10 is a diagram illustrating an example of the contents of a data table. [Figure 6] FIG. 10 is a diagram illustrating an example of the contents of a data table. [Figure 7] FIG. 10 is a diagram illustrating an example of the contents of a data table. [Figure 8] FIG. 10 is a diagram illustrating an example of the contents of a data table. [Figure 9] FIG. 10 is a diagram illustrating an example of the contents of a data table. [Figure 10] FIG. 10 is a diagram illustrating an example of the contents of a data table. [Figure 11] FIG. 10 is a diagram illustrating an example of the characteristics of the overall prediction accuracy with respect to the amount of data complementation measurement work. DETAILED DESCRIPTION OF THE INVENTION

[0010] Hereinafter, an embodiment of the present invention will be described in detail with reference to the drawings.

[0011] First, the gist of this embodiment will be explained. In recent years, highly accurate risk prediction models have been developed using a wide variety of data. However, when combining and inputting such data into a risk prediction model, the presence of missing data can pose a problem. While information collection to compensate for such missing data is sometimes performed individually, such information collection requires a significant amount of work. On the other hand, to achieve the target prediction accuracy in a risk prediction model, it is not necessary to collect information to compensate for all missing data. For example, if 20% of the entries (hereinafter simply referred to as "entries") are related to 80% of the risks, compensating for the defective data of only these 20% of the entries is sufficient to achieve the target predictability. Therefore, in this embodiment, a compensation priority flag is assigned to entries estimated to be related to 80% of the risks, and an "integrated compensation priority flag" is assigned, which indicates a ranking according to the number of entries to which a compensation priority flag is assigned, i.e., the number of compensation priority flags. Using these flags, this embodiment achieves missing data compensation that contributes to high prediction accuracy with less work. A detailed description will be given below.

[0012] 1 is a system configuration diagram showing an example of the configuration of a missing data impregnation device 100 according to the first embodiment. The missing data impregnation device 100 is, for example, a computer, and includes a data table 10, a parameter definition unit 20, a missing data prediction unit 30, an extraction unit 40, a impregnation priority flag assignment unit 50, and a priority ranking determination unit 60, and preferably includes a display unit 70. The display unit 70 may be integrated with the missing data impregnation device 100 or may be separate from it.

[0013] The data table 10 has a plurality of column data items defined in the column direction and a plurality of entries each consisting of data for a plurality of column data items in the row direction. In the data table 10, some of the entries have missing data in specific column data items (flood risk, exchange risk).

[0014] The parameter definition unit 20 defines a complement amount adjustment parameter X, and this complement amount adjustment parameter C is a parameter used to adjust the proportion of data that needs to be complemented for a specific column data item (for example, "flood risk" and "exchange risk" described below). In this embodiment, a predetermined proportion, for example, the top 30%, is defined as this complement amount adjustment parameter X, and preferably, the top 70% is defined.

[0015] The missing data prediction unit 30 predicts missing data for a specific column data item based on data for other column data items other than the specific column data item and the data for the specific column data item in multiple entries that make up the data table 10.

[0016] The missing data prediction unit 30 learns a model using, for example, data of column data items other than a specific column data item as explanatory variables and data of the specific column data item as a target variable in a plurality of entries that make up the data table 10. The missing data prediction unit 30 predicts missing data of the specific column data item by machine learning using this model.

[0017] The extraction unit 40 extracts at least one entry that is included in a predetermined ratio defined by the complement amount adjustment parameter X from each piece of data in a specific column data item.

[0018] The complementary priority flag assigning unit 50 assigns a complementary priority flag to at least one entry in the data table 10 that falls within the predetermined ratio defined by the complementary amount adjustment parameter X as described above.

[0019] The priority order determining unit 60 counts the number of entries included in the data table 10 to which complementary priority flags have been assigned, and determines the integrated complementary priority order in descending order of the number of complementary priority flags assigned.

[0020] The display unit 70 displays the data for each specific column data item and the list data of the integration and completion priority order for each entry that includes missing data. The contents of the display will be described later.

[0021] The complementation priority flag assigning unit 50 assigns a complementation priority flag for each of the plurality of predetermined ratios (e.g., 30%, 70%) defined as the complementation amount adjustment parameters. The display unit 70 displays the amount of data complementation measurement work based on the above-mentioned list data and the prediction accuracy of the missing data prediction unit 30 for data of specific column data items (e.g., "flood risk" and "exchange risk" described below).

[0022] The missing data imputation device 100 is configured as described above, and next, a missing data imputation method will be described as an example of the operation of the missing data imputation device 100. The missing data imputation method includes a parameter definition step in which a parameter definition unit defines an imputation amount adjustment parameter X used to adjust the proportion of data that needs to be imputed for a specific column data item in a data table 10 in which a plurality of column data items are defined in the column direction and a plurality of entries each made up of data for a plurality of column data items in the row direction, and in which missing data exists in a specific column data item in some of the entries; the missing data prediction step of predicting missing data in a specific column data item; the extraction step in which the extraction unit 40 extracts at least one entry included in a predetermined ratio defined by the complement amount adjustment parameter X from each of the data in the specific column data item; the complement priority flag assignment step in which the complement priority flag assignment unit 50 assigns a complement priority flag to at least one entry included in the predetermined ratio; and the priority ranking determination step in which the priority ranking determination unit 60 counts the number of entries included in the data table 10 to which a complement priority flag has been assigned, and determines an integrated complement priority ranking in descending order of the number of complement priority flags assigned.

[0023] Fig. 2 is a flowchart showing an example of the procedure for missing data complementation processing, and Fig. 3 is a flowchart showing an example of the procedure for missing data processing (step S20 in Fig. 2) shown in Fig. 2. The following description will be given with reference to data table 10 shown in Figs. 4 to 9.

[0024] 2, the data table 10 and the compensation amount adjustment parameter X are input to the missing data compensation device 100. The parameter definition unit 20 sets the compensation amount adjustment parameter X to, for example, the top 30% based on the inputted parameter.

[0025] On the other hand, as shown in Figure 4, data table 10 defines, as column data items, for example, business establishment ID (IDentifier), country of location, long-term interest rate, distance from river, flood risk, flood-related supplementary priority, exchange rate risk, exchange rate-related supplementary priority, and overall supplementary priority. Data table 10 is made up of data for the column data items and manages entries that are mutually identified by the business establishment ID (IDentifier) ​​described above. In data table 10 shown in Figure 4, for example, there is missing data indicated by "-" in flood risk and exchange rate risk, which are examples of specific column data items.

[0026] In step S20, missing data processing is performed. The missing data processing is performed by the missing data prediction unit 30. In step S21 of FIG. 3, the missing data prediction unit 30 predicts the data of the specific column data item (missing value indicated by "-" in the drawing).

[0027] Specifically, for example, with regard to flood risk, the missing data prediction unit 30 creates a model that predicts a specific column data item containing missing data, such as flood risk data, from a combination of entries that do not contain missing data. The combination of entries that do not contain missing data here refers, for example, to the entries with business IDs "AAAA001," "AABB002," "DDFF004," "AAFF005," and "CCGG006" shown in bold frame 10A in FIG. 5. In this embodiment, flood risk = S is affected by the country of location and the distance from the river. In this embodiment, these two variables (country of location and distance from the river) are used as explanatory variables, and flood risk is used as the target variable.

[0028] Meanwhile, for example, with regard to exchange risk, the missing data prediction unit 30 creates a model that predicts specific column data items containing missing data, such as exchange risk data, from combinations of entries that do not contain missing data. The combination of entries that do not contain missing data here refers, for example, to the entries with business IDs "AAAA002," "CCDD003," "DDFF004," and "CCGG006" shown in bold frame 10B in FIG. 6. In this embodiment, exchange risk = K is assumed to be affected by the long-term interest rate of the country of domicile. In this embodiment, this (the long-term interest rate of the country of domicile) is used as the explanatory variable, and exchange risk is used as the target variable.

[0029] The missing data prediction unit 30, for example, learns a model using data of column data items other than a specific column data item in multiple entries that make up the data table 10 as explanatory variables and data of a specific column data item (e.g., flood risk, exchange rate risk) as a target variable, and predicts missing data of a specific column data item using machine learning using this model (corresponding to the "(predicted)" part shown in Figure 7).

[0030] In step S22 of FIG. 3, the extraction unit 40 extracts at least one entry from each data item of a specific column data item (e.g., flood risk, exchange risk) that is included in a predetermined proportion (e.g., the top 30%) defined by the complement amount adjustment parameter X.

[0031] 3, the complementary priority flag assigning unit 50 assigns a complementary priority flag to at least one entry included in a predetermined percentage (for example, the top 30%) defined by the complementary amount adjustment parameter X (see FIG. 8). In the illustrated example, entries to which a complementary priority flag has been assigned are marked with a double circle.

[0032] Next, in step S30 of FIG. 2, the priority ranking determination unit 60 counts the number of entries to which complementary priority flags are assigned for each entry included in the data table 10, and determines the integrated complementary priority ranking in descending order of the number of complementary priority flags (see FIG. 9).

[0033] In step S40, the display unit 70 displays the list data of entries with missing data x (data for each column data item, integration and complementation priority order).

[0034] In step S50, the display unit 70 displays pairs of the complementary man-hour amount and the overall prediction accuracy for a plurality of complementary amount adjustment parameters X.

[0035] 10 is a diagram showing an example of a case where missing data complementation processing is performed with the complement amount adjustment parameter X=70%. In the example shown in FIG. 8, the same content as in the examples shown in FIGS. 8 and 9 will not be described.

[0036] In the example shown in FIG. 10, the complement amount adjustment parameter X=70%, and the extraction unit 40 extracts at least one entry that is included in the top 70% of each data item of a specific column data item (for example, flood risk, exchange risk).

[0037] In addition to the data 10D extracted in Figures 8 and 9 described above, the supplementary priority flag assigning unit 50 also assigns a supplementary priority flag to the flood risk data 10E of the entry for the business ID "CCDD003" (corresponding to the "◎" supplementary priority for flood risk shown in the figure).

[0038] The priority ranking determination unit 60 counts the number of entries included in the data table 10 to which a complementary priority flag has been assigned, and determines the integrated complementary priority ranking in descending order of the number of complementary priority flags. As a result, the integrated complementary priority ranking is such that the entry with the business establishment ID "SSDD007" with two "◎" marks is ranked first, the entry with the business establishment ID "CCDD003" with one "◎" marks is ranked second, and the entry with the business establishment ID "AAFF005" with one "◎" marks is ranked third. Therefore, the example shown in FIG. 10 differs from the example shown in FIG. 9 in that the entry with the business establishment ID "CCDD003" is also included.

[0039] 11 is a diagram showing an example of the characteristics of overall prediction accuracy versus data complement measurement effort. The illustrated example shows a case where the complement amount adjustment parameter X is 70% and a case where the complement amount adjustment parameter X is 30%, with the target prediction accuracy (corresponding to the "target accuracy" shown) set to 0.75. In the illustrated example, "1ID" indicates complementation using an entry corresponding to one establishment ID (hereinafter also referred to as "1ID complementation"), "2ID" indicates complementation using entries corresponding to two establishment IDs (hereinafter also referred to as "2ID complementation"), and "3ID" indicates complementation using entries corresponding to three establishment IDs (hereinafter also referred to as "3ID complementation").

[0040] Here, the number of entries corresponding to establishment IDs to be complemented is determined while taking into consideration the target accuracy described above. In this embodiment, the missing data prediction unit 30 considers how to set the value of the complement amount adjustment parameter X (here, either 30% or 70%) to obtain a prediction accuracy of "0.75" as an example of the target accuracy. That is, the missing data prediction unit 30 selects one predetermined percentage from multiple predetermined percentages (e.g., 30% or 70%) defined as the complement amount adjustment parameter X so that the overall prediction accuracy of the training data including missing data satisfies the target accuracy.

[0041] In the example shown, the amount of data supplementary measurement work increases in the order of "1ID," "2ID," and "3ID," and the overall prediction accuracy improves accordingly. The prediction accuracy of "2ID" is, for example, 0.76, the prediction accuracy of "1ID" is less than 0.75 (for example, 0.67), and the prediction accuracy of "3ID" is greater than 0.75 (for example, 0.82).

[0042] In this embodiment, when the complement amount adjustment parameter X=30%, if the prediction accuracy is to be ensured to ensure the above-mentioned target accuracy of "0.75", the prediction accuracy is insufficient with 1 ID complementation, so 2 ID complementation is used.

[0043] On the other hand, in this embodiment, when the complement amount adjustment parameter X=70%, if we try to achieve a prediction accuracy that ensures the above-mentioned target accuracy of "0.75", the prediction accuracy is insufficient with 1 ID complementation, so 3 ID complementation is used.

[0044] As described above, by setting the interpolation amount adjustment parameter X so as to achieve a prediction accuracy corresponding to the target accuracy of 0.75 described above, the missing data prediction unit 30 can improve the overall prediction accuracy of the training data including missing data using the above-described model.

[0045] As described above, the missing data imputation device 100 according to this embodiment includes: a data table 10 having a plurality of column data items defined in the column direction and a plurality of entries in the row direction, each of which is composed of data for the plurality of column data items, and in which missing data exists for a specific column data item (e.g., flood risk, exchange rate risk) in some of the entries; a parameter definition unit 20 that defines a imputation amount adjustment parameter X used to adjust the proportion of data in the specific column data item that needs to be imputed; a missing data prediction unit 30 that predicts missing data for a specific column data item based on data in column data items other than the specific column data item and the data in the specific column data item in the plurality of entries that make up the data table 10; an extraction unit 40 that extracts at least one entry that falls within a predetermined proportion (e.g., 30%) defined by the imputation amount adjustment parameter X; a imputation priority flag assignment unit 50 that assigns a imputation priority flag to at least one entry that falls within the predetermined proportion; and a priority ranking determination unit 60 that counts the number of entries in the data table 10 to which a imputation priority flag has been assigned, and determines an integrated imputation priority ranking in descending order of the number of imputation priority flags assigned.

[0046] In this way, the extraction unit 40 extracts at least one entry that falls within the predetermined ratio defined by the completion amount adjustment parameter X, assigns a completion priority flag to this entry, and determines the integrated completion priority ranking in descending order of the number of completion priority flags. Therefore, by referring to the integrated completion priority ranking, it is possible to accurately grasp the missing data that has a significant impact on prediction accuracy, and therefore it is possible to improve the overall prediction accuracy of the training data that includes missing data while minimizing the amount of missing data that needs to be completed.

[0047] The missing data imputation device 100 according to this embodiment includes a display unit 70 that displays, for each entry containing missing data, the data for each specific column data item and the list data of the integration and imputation priorities. In this way, by referring to the data for each specific column data item and the list data of the integration and imputation priorities for each entry containing missing data, it is possible to improve the overall prediction accuracy of the training data containing missing data while minimizing the amount of missing data to be imputed.

[0048] In this embodiment, the complement priority flag assigning unit 50 assigns a complement priority flag for each of a plurality of predetermined ratios (e.g., 30%, 70%) defined as the complement amount adjustment parameter X, and the display unit 70 displays the amount of data complement measurement work based on the list data and the prediction accuracy for data of a specific column data item by the missing data prediction unit 30. In this way, by referring to the amount of data complement measurement work based on the list data and the prediction accuracy for data of a specific column data item by the missing data prediction unit 30, it is possible to reduce the amount of missing data to be complemented while improving the overall prediction accuracy of training data including missing data.

[0049] In this embodiment, the missing data prediction unit 30 selects one predetermined ratio from multiple predetermined ratios defined as the complement amount adjustment parameter X so that the overall prediction accuracy of the training data including missing data satisfies the target accuracy. In this way, it is possible to increase the overall prediction accuracy of the training data including missing data to above the target accuracy while minimizing the amount of missing data to be complemented.

[0050] In this embodiment, the missing data prediction unit 30 learns a model using data of column data items other than the specific column data item in the multiple entries that make up the data table 10 as explanatory variables and data of the specific column data item (for example, the above-mentioned flood risk and exchange rate risk) as a response variable, and predicts missing data of the specific column data item using the model through machine learning. In this way, the missing data prediction unit 30 can predict missing data using the model while minimizing the amount of missing data to be complemented and improving the overall prediction accuracy of the training data that includes missing data.

[0051] The present invention is not limited to the above-described embodiments, and includes various modifications and equivalent configurations within the spirit and scope of the appended claims. For example, the above-described embodiments have been described in detail to clearly explain the present invention, and the present invention is not necessarily limited to those including all of the described configurations. Furthermore, the elements described in parallel in the present embodiment may be configured such that at least one of the elements is connected in series to the other elements. [Industrial Applicability]

[0052] The present invention can be applied to a missing data filling device relating to a technique for filling missing data. [Explanation of symbols]

[0053] 10...data table, 20...parameter definition unit, 30...missing data prediction unit, 40...extraction unit, 50...complementation priority flag assignment unit, 60...priority ranking determination unit, 100...missing data complementation device

Claims

1. a data table in which a plurality of column data items are defined in the column direction and a plurality of entries each consisting of data of the plurality of column data items in the row direction, wherein missing data exists in a specific column data item in some of the entries; a parameter definition section for defining an interpolation amount adjustment parameter used to adjust the proportion of data in the specific column data item that needs to be interpolated; a missing data prediction unit that predicts the missing data of the specific column data item based on data of other column data items other than the specific column data item in the plurality of entries that make up the data table, and the data of the specific column data item; an extracting unit that extracts at least one entry included in a predetermined ratio defined by the complement amount adjustment parameter from each data item of the specific column data item; a complementary priority flag assigning unit that assigns a complementary priority flag to at least one of the entries included in the predetermined ratio; a priority ranking determination unit that counts the number of entries to which the complementary priority flags are assigned for each of the entries included in the data table, and determines an integrated complementary priority ranking in descending order of the number of complementary priority flags; A missing data compensation device comprising:

2. a display unit that displays, for each entry including the missing data, the data for each specific column data item and the list data of the integration and completion priority order; 2. The missing data compensation device according to claim 1.

3. The complementary priority flag assigning unit assigning the complementation priority flag to each of the plurality of predetermined ratios defined as the complementation amount adjustment parameter; The display unit The amount of data supplementation measurement work based on the list data and the prediction accuracy of the missing data prediction unit for the data of the specific column data item are displayed.

3. The missing data compensation device according to claim 2.

4. The missing data prediction unit Selecting one predetermined ratio from the plurality of predetermined ratios defined as the complement amount adjustment parameter so that the overall prediction accuracy of the training data including the missing data satisfies the target accuracy.

4. The missing data compensation device according to claim 3.

5. The missing data prediction unit A model is trained using data of column data items other than the specific column data item in the plurality of entries constituting the data table as explanatory variables and data of the specific column data item as a target variable, and the missing data of the specific column data item is predicted by machine learning using the model.

2. The missing data compensation device according to claim 1.

6. a parameter definition step of defining in a parameter definition section an interpolation amount adjustment parameter used to adjust the proportion of data in a specific column data item in a data table in which a plurality of column data items are defined in a column direction and a plurality of entries are formed from each piece of data in the plurality of column data items in a row direction, and in which missing data exists in a specific column data item in some of the entries; a missing data prediction step in which a missing data prediction unit predicts the missing data of the specific column data item based on data of column data items other than the specific column data item and the data of the specific column data item in the plurality of entries constituting the data table; an extraction step in which an extraction unit extracts at least one entry included in a predetermined ratio defined by the complement amount adjustment parameter from each data item of the specific column data item; a complementary priority flag assigning step in which a complementary priority flag assigning unit assigns a complementary priority flag to at least one of the entries included in the predetermined ratio; a priority order determination step in which a priority order determination unit counts the number of entries to which the complementary priority flags are assigned for each of the entries included in the data table, and determines an integrated complementary priority order in descending order of the number of complementary priority flags; A missing data imputation method comprising:

7. a display step of displaying, on a display unit, the data for each specific column data item and the list data of the integration and completion priority order for each entry including the missing data; 7. The missing data filling method according to claim 6.

8. The complementary priority flag assigning unit assigning the complementation priority flag to each of a plurality of values ​​set as the complementation amount adjustment parameter; In the display step, The display unit displays the amount of data complementation measurement work based on the list data and the prediction accuracy of the missing data prediction unit for the data of the specific column data item.

8. The missing data filling method according to claim 7.

9. In the missing data prediction step, The missing data prediction unit selects one predetermined ratio from the plurality of predetermined ratios defined as the complement amount adjustment parameter so that the overall prediction accuracy of the training data including the missing data satisfies a target accuracy.

9. The missing data filling method according to claim 8.

10. In the missing data prediction step, The missing data prediction unit learns a model using data of column data items other than the specific column data item in the plurality of entries constituting the data table as explanatory variables and data of the specific column data item as a target variable, and predicts the missing data of the specific column data item by machine learning using the model.

7. The missing data filling method according to claim 6.

Images