Detailed description of computation types

Article Summary

Share feedback

Thanks for sharing your feedback!

Available in VPC

This document describes the different computation features you can select when creating a data computation task according to the instructions in Data computation in detail, and guides you through setting them.

Manage columns

Copy, move, or extract a specified column from the computed data to result data, or delete it from the computed data.

Copy column

Select data and column from Computed data and Target column respectively, and then select Result data to copy the column to.

Copy the selected target column and add to Result data.
You can select only 1 target column.
The copying column computation will fail if you select the same value for Computed data and Result data.

Return to the information about task creation

Move column

Select data and column from Computed data and Target column respectively, and then select Result data to move the column to.

Move the selected target column to Result data. When the moving is completed, the selected target column is deleted from Input data.
You can select only 1 target column.
The copying column computation will fail if you select the same value for Computed data and Result data.

Return to the information about task creation

Extract column

Select data and column from Computed data and Target column respectively.

Extract the selected target column, and then add it to the newly created result data. When the extraction is completed, the selected target column is deleted from Computed data.
When the extraction and addition are complete, a {computed data} _ export ciphertext is created that includes the {column name} column.
Multiple target columns can be selected, and all selected target columns will be extracted.

Return to the information about task creation

Delete column

Select data and column from Computed data and Target column respectively.

Delete the selected target column from the Computed data.
You can select only 1 target column.

Return to the information about task creation

Statistical analysis

Use the value in the column specified in the computation data to perform various statistics computations such as addition, average, and distribution.

Addition between columns

Select two individual columns from the Computed data, add their values, and then save the result in the result column.

The following describes how to enter detailed settings.

Select Computed data.
Individually select Column 1 and Column 2 to perform addition.
- Only columns that use real number as the value type can be selected.
- Addition between the same column is also possible.
Click the [Add] button on the right side of the Result column to add the addition computation condition you've just entered. Repeat the procedure to add all addition processes you want performed.

Return to the information about task creation

Subtraction between columns

Select two individual columns from the Computed data, calculate their difference, and then save the result in the result column.

The following describes how to enter detailed settings.

Select Computed data.
Individually select Column 1 and Column 2 to perform subtraction.
- Only columns that use real number as the value type can be selected.
- Subtraction between the same column is also possible.
Click the [Add] button on the right side of the Result column to add the subtraction computation condition you've just entered. Repeat the procedure to add all the subtraction processes you want performed.

Return to the information about task creation

Multiplication between columns

Select two individual columns from the Computed data, multiply the values, and then save the result in the result column.

The following describes how to enter detailed settings.

Select Computed data.
Individually select Column 1 and Column 2 to perform multiplication.
- Only columns that use real number as the value type can be selected.
- Multiplication between the same column is also possible.
Click the [Add] button on the right side of the Result column to add the multiplication computation condition you've just entered. Repeat the procedure to add all multiplication processes you want performed.

Return to the information about task creation

Addition between column and constant

Add a designated value for the constant of the selected column from the Computed data, and then save the result in the result column.

The following describes how to enter detailed settings.

Select Computed data.
Select the Column to perform addition, and then enter a Constant to add to the column.
- Only columns that use real number as the value type can be selected.
Click the [Add] button on the right side of the Result column to add the addition computation condition you've just entered. Repeat the procedure to add all addition processes you want performed.

Return to the information about task creation

Subtraction between column and constant

Subtract a designated value for the constant of the selected column from the Computed data, and then save the result in the result column.

The following describes how to enter detailed settings.

Select Computed data.
Select the Column to perform subtraction, and then enter a Constant to subtract from the column.
- Only columns that use real number as the value type can be selected.
Click the [Add] button on the right side of the Result column to add the subtraction computation condition you've just entered. Repeat the procedure to add all the subtraction processes you want performed.

Return to the information about task creation

Multiplication between column and constant

Multiply a designated value for the constant to the selected column from the Computed data, and then save the result in the result column.

The following describes how to enter detailed settings.

Select Computed data.
Select the Column to perform multiplication, and then enter a Constant to multiply by the column.
- Only columns that use real number as the value type can be selected.
Click the [Add] button on the right side of the Result column to add the multiplication computation condition you've just entered. Repeat the procedure to add all multiplication processes you want performed.

Return to the information about task creation

Sum

Select the Computed data, and then select the column to calculate the sum of in Target column.

All values in the selected Target column will be added up.
You can select only 1 Target column.
A column to save the results is created automatically, but you can specify a column name you want on the Result column.
Select the Save result column separately to create new data other than Computed data to save. You can enter the desired data name on the Result data.

Return to the information about task creation

Average

Select the Computed data, and then select the column to calculate the average in Target column.

The average of the values in the selected Target column will be calculated.
You can select only 1 Target column.
A column to save the results is created automatically, but you can specify a column name you want on the Result column.
Select the Save result column separately to create new data other than Computed data to save. You can enter the desired data name on the Result data.

Return to the information about task creation

Distribution

Select the Computed data, and then select the column to calculate the distribution for in Target column.

The distribution of the values in the selected Target column will be calculated.
You can select only 1 Target column.
A column to save the results is created automatically, but you can specify a column name you want on the Result column.
Select the Save result column separately to create new data other than Computed data to save. You can enter the desired data name on the Result data.

Return to the information about task creation

Standard deviation

Select the Computed data, and then select the column to calculate the standard deviation for in Target column.

The standard deviation of the values in the selected Target column will be calculated.
You can select only 1 Target column.
A column to save the results is created automatically, but you can specify a column name you want on the Result column.
Select the Save result column separately to create new data other than Computed data to save. You can enter the desired data name on the Result data.

Return to the information about task creation

Standard error

Select the Computed data, and then select the column to calculate the standard error for in Target column.

The standard error of the values in the selected Target column will be calculated.
You can select only 1 Target column.
A column to save the results is created automatically, but you can specify a column name you want on the Result column.
Select the Save result column separately to create new data other than Computed data to save. You can enter the desired data name on the Result data.

Return to the information about task creation

Coefficient of variation

Select the Computed data, and then select the column to calculate the coefficient of variation for in Target column.

The coefficient of variation of the values in the selected Target column will be calculated.
You can select only 1 Target column.
A column to save the results is created automatically, but you can specify a column name you want on the Result column.
Select the Save result column separately to create new data other than Computed data to save. You can enter the desired data name on the Result data.

Return to the information about task creation

Absolute value

Select the Computed data, and then select the column to calculate the absolute value for in Target column.

The coefficient of variation of the values in the selected Target column will be calculated.
You can select only 1 Target column.
A column to save the results is created automatically, but you can specify a column name you want on the Result column.
Select the Save result column separately to create new data other than Computed data to save. You can enter the desired data name on the Result data.

Return to the information about task creation

Skewness

Select the Computed data, and then select the column to calculate the skewness for in Target column.

The coefficient of variation of the values in the selected Target column will be calculated.
You can select only 1 Target column.
A column to save the results is created automatically, but you can specify a column name you want on the Result column.
Select the Save result column separately to create new data other than Computed data to save. You can enter the desired data name on the Result data.

Return to the information about task creation

Kurtosis

Select the Computed data, and then select the column to calculate the kurtosis for in Target column.

The coefficient of variation of the values in the selected Target column will be calculated.
You can select only 1 Target column.
A column to save the results is created automatically, but you can specify a column name you want on the Result column.
Select the Save result column separately to create new data other than Computed data to save. You can enter the desired data name on the Result data.

Return to the information about task creation

Correlation

Select the computational data, select two columns for correlation in the target column, and click the Add Condition button.

Find the correlation between the two selected columns.
Columns that store result values are automatically created, but you can name the column you want in the resulting column.
The result column creates and stores new data, and you can enter the desired data name into the result data. If not entered, the name of the result data is automatically generated by referring to the input data.

Return to the information about task creation

Covariance

Select the computational data, select two columns for covariance in the target column, and click the Add Condition button.

Find the covariance between the two selected columns.
Columns that store result values are automatically created, but you can name the column you want in the resulting column.
The result column creates and stores new data, and you can enter the desired data name into the result data. If not entered, the name of the result data is automatically generated by referring to the input data.

Return to the information about task creation

Filter by category

Filter the cells that meet certain criteria among the cells in the selected column in the Computed data, and then save the result in a new column.

The following describes how to enter detailed settings.

Select Computed data.
Select the target column to filter in Target column.
- Only columns that use real number as the value type can be selected.
From the Computation condition area, select Category column and Category value which will serve as the conditions, and then click the [Add] button on the right. Repeat the procedure to add all target cell values you want.
- Only columns with the category as value type can be selected.

A column to save the results is created automatically, but you can specify a column name you want on the Result column.
Select the Save result column separately to create new data other than Computed data to save. You can enter the desired data name on the Result data.
Return to the information about task creation

Frequency by category

Count the number of cells that meet certain criteria in the selected column in the Computed data, and then save the value in a new column.

The following describes how to enter detailed settings.

Select Computed data.
Select the target column to count the frequency in Target column.
- Only columns that use real number as the value type can be selected.
From the Computation condition area, select Category column and Category value which will serve as the conditions, and then click the [Add] button on the right. Repeat the procedure to add all target cell values you want.
- Only columns with the category as value type can be selected.

A column to save the results is created automatically, but you can specify a column name you want on the Result column.
Select the Save result column separately to create new data other than Computed data to save. You can enter the desired data name on the Result data.
Return to the information about task creation

Merge class

Merge a specific cell value in the selected column in the Computed data with a new cell value, and then save the result in a new column.

The following describes how to enter detailed settings.

Select Computed data.
Select the target column to merge in Target column.
- Only columns that use the category number as the value type can be selected.
Enter the Existing cell value that will be merged.
- Use commas (,) as separators if you want to enter multiple values.
Enter the New cell value.

A column to save the results is created automatically, but you can specify a column name you want on the Result column.
Select the Save result column separately to create new data other than Computed data to save. You can enter the desired data name on the Result data.
Return to the information about task creation

Sum by category

Calculate the sum of all cells that meet certain criteria in the selected column in the Computed data, and then save the result in a new column.

The following describes how to enter detailed settings.

Select Computed data.
Select the target column to aggregate in Target column.
- Only columns that use real number as the value type can be selected.
From the Computation condition area, select Category column and Category value which will serve as the conditions, and then click the [Add] button on the right. Repeat the procedure to add all target cell values you want.
- Only columns with the category as value type can be selected.

A column to save the results is created automatically, but you can specify a column name you want on the Result column.
Select the Save result column separately to create new data other than Computed data to save. You can enter the desired data name on the Result data.
Return to the information about task creation

Average by category

Calculate the average of all cells that meet certain criteria in the selected column in the Computed data, and then save the result in a new column.

The following describes how to enter detailed settings.

Select Computed data.
Select the target column to calculate the average in Target column.
- Only columns that use real number as the value type can be selected.
From the Computation condition area, select Category column and Category value which will serve as the conditions, and then click the [Add] button on the right. Repeat the procedure to add all target cell values you want.
- Only columns with the category as value type can be selected.

A column to save the results is created automatically, but you can specify a column name you want on the Result column.
Select the Save result column separately to create new data other than Computed data to save. You can enter the desired data name on the Result data.
Return to the information about task creation

Distribution by category

Calculate the distribution of all cells that meet certain criteria in the selected column in the Computed data, and then save the result in a new column.

The following describes how to enter detailed settings.

Select Computed data.
Select the target column to calculate the distribution in Target column.
- Only columns that use real number as the value type can be selected.
From the Computation condition area, select Category column and Category value which will serve as the conditions, and then click the [Add] button on the right. Repeat the procedure to add all target cell values you want.
- Only columns with the category as value type can be selected.

A column to save the results is created automatically, but you can specify a column name you want on the Result column.
Select the Save result column separately to create new data other than Computed data to save. You can enter the desired data name on the Result data.
Return to the information about task creation

Standard deviation by category

Calculate the standard deviation of all cells that meet certain criteria in the selected column in the Computed data, and then save the result in a new column.

The following describes how to enter detailed settings.

Select Computed data.
Select the target column to calculate the standard deviation in Target column.
- Only columns that use real number as the value type can be selected.
From the Computation condition area, select Category column and Category value which will serve as the conditions, and then click the [Add] button on the right. Repeat the procedure to add all target cell values you want.
- Only columns with the category as value type can be selected.

A column to save the results is created automatically, but you can specify a column name you want on the Result column.
Select the Save result column separately to create new data other than Computed data to save. You can enter the desired data name on the Result data.
Return to the information about task creation

Standard error by category

Calculate the standard error of all cells that meet certain criteria in the selected column in the Computed data, and then save the result in a new column.

The following describes how to enter detailed settings.

Select Computed data.
Select the target column to calculate the standard error in Target column.
- Only columns that use real number as the value type can be selected.
From the Computation condition area, select Category column and Category value which will serve as the conditions, and then click the [Add] button on the right. Repeat the procedure to add all target cell values you want.
- Only columns with the category as value type can be selected.

A column to save the results is created automatically, but you can specify a column name you want on the Result column.
Select the Save result column separately to create new data other than Computed data to save. You can enter the desired data name on the Result data.
Return to the information about task creation

Coefficient of variation by category

Calculate the coefficient of variation of all cells that meet certain criteria in the selected column in the Computed data, and then save the result in a new column.

The following describes how to enter detailed settings.

Select Computed data.
Select the target column to calculate the coefficient of variation in Target column.
- Only columns that use real number as the value type can be selected.
From the Computation condition area, select Category column and Category value which will serve as the conditions, and then click the [Add] button on the right. Repeat the procedure to add all target cell values you want.
- Only columns with the category as value type can be selected.

A column to save the results is created automatically, but you can specify a column name you want on the Result column.
Select the Save result column separately to create new data other than Computed data to save. You can enter the desired data name on the Result data.
Return to the information about task creation

Absolute value by category

Find the absolute value of all cell values satisfying the conditions in the selected column of the computational data and store the results in a new column.

The detailed way to enter the detailed settings is as follows.

Please select computational data.
Please select a target column to obtain the absolute values from the target column.
- Only columns that use real number as the value type can be selected.
In the Computational Conditions area, select the category column and category value that will be the condition and click the Add button on the right. Please repeat this procedure to add all the desired target cell values.
- Only columns with categorical value types can be selected.

Columns that store result values are automatically created, but you can name the column you want in the resulting column.
The result column creates and stores new data, and you can enter the desired data name into the result data. If not entered, the name of the result data is automatically generated by referring to the input data.
Return to the information about task creation

Skewness by category

Find the skewness of all cell values satisfying the conditions in the selected column of the computational data and store the results in a new column.

The detailed way to enter the detailed settings is as follows.

Please select computational data.
Please select a target column to obtain the skewness from the target column.
- Only columns that use real number as the value type can be selected.
In the Computational Conditions area, select the category column and category value that will be the condition and click the Add button on the right. Please repeat this procedure to add all the desired target cell values.
- Only columns with categorical value types can be selected.

Columns that store result values are automatically created, but you can name the column you want in the resulting column.
The result column creates and stores new data, and you can enter the desired data name into the result data. If not entered, the name of the result data is automatically generated by referring to the input data.
Return to the information about task creation

Kurtosis by category

Find the kurtosis of all cell values satisfying the conditions in the selected column of the computational data and store the results in a new column.

The detailed way to enter the detailed settings is as follows.

Please select computational data.
Please select a target column to obtain the kurtosis from the target column.
- Only columns that use real number as the value type can be selected.
In the Computational Conditions area, select the category column and category value that will be the condition and click the Add button on the right. Please repeat this procedure to add all the desired target cell values.
- Only columns with categorical value types can be selected.

Columns that store result values are automatically created, but you can name the column you want in the resulting column.
The result column creates and stores new data, and you can enter the desired data name into the result data. If not entered, the name of the result data is automatically generated by referring to the input data.
Return to the information about task creation

Maximum value

Select the Computed data, and then select the column to calculate the maximum value for in Target column.

The coefficient of variation of the values in the selected Target column will be calculated.
You can select only 1 Target column.
A column to save the results is created automatically, but you can specify a column name you want on the Result column.
Select the Save result column separately to create new data other than Computed data to save. You can enter the desired data name on the Result data.

Return to the information about task creation

Minimum value

Select the Computed data, and then select the column to calculate the minimum value for in Target column.

The coefficient of variation of the values in the selected Target column will be calculated.
You can select only 1 Target column.
A column to save the results is created automatically, but you can specify a column name you want on the Result column.
Select the Save result column separately to create new data other than Computed data to save. You can enter the desired data name on the Result data.

Return to the information about task creation

Range

Select the Computed data, and then select the column to calculate the range for in Target column.

The coefficient of variation of the values in the selected Target column will be calculated.
You can select only 1 Target column.
A column to save the results is created automatically, but you can specify a column name you want on the Result column.
Select the Save result column separately to create new data other than Computed data to save. You can enter the desired data name on the Result data.

Return to the information about task creation

Machine learning

You can use the selected data to perform logistic regression learning, create inference results, and perform decryption.

Learning - logistic regression

Use the selected Data for learning to perform logistic regression learning, and create encrypted logistic regression coefficient tables.

A logistic regression coefficient table, as a logistic regression model with its training completed, can be used as Inference model data when you use the Inference - logistic regression feature.

The following describes how to enter detailed settings.

Select the Data for learning.
Enter the Learning rate value larger than 0.
Enter Epochs.
Enter the value of Mini batch.
- The minimum value is 1, and the maximum value is determined according to the encrypted key of the Data for learning and the number of columns and rows in the data. If a number greater than the maximum value is entered, the maximum value will automatically be adjusted.

Return to the information about task creation

Inference - logistic regression

Select Inference model, and then select Inference data to perform inference.

Apply the inference model to the selected Inference data to perform inference, and create an encrypted inference result data.
The inference result data can be used as Computed data when using the Decrypt inference result feature.

Select inference model

You can use a logistic regression coefficient table created through the Learning - logistic regression feature, or an independent plaintext inference model file as an Inference model.

To upload a plaintext inference model file, click the [Upload plaintext model] button, enter File type, Class count in the upload pop-up window, upload the file, and then click the [Confirm] button.

Return to the information about task creation

Decrypt inference result

Select the target Computed data to decrypt.

The selected Computed data will be decrypted to create plaintext data of the inference result.
You can select the Computed data from among the result data created through the Inference - logistic regression feature.

Return to the information about task creation

Was this article helpful?

What's Next

Using agent

Table of contents

Manage columns
Statistical analysis
Machine learning