Help Manual

Contents






Sigma Magic Help Version 17

Make Predictions

Overview

The Analytics models you have developed can be used to predict new data sets. For example, let's say you have created a model to predict high-risk and low-risk customers for giving out loans. Based on some historical data you have collected, you have information on factors such as age, education level, the amount of loan taken, salary, family status, etc, as input factors that may impact the risk. You can develop a model using one of the many available analytics models based on this historical data. Once this model is created, you want to use it in your practice to see if a new customer walks in the door and if they are high-risk or low-risk customers. You can use the developed model to make this prediction. The Make Predictions module can be used for the following analytics models.
NumModel Name
1Bagging Models
2Bayesian Models
3Boosted Models
4Decision Trees
5Discriminant Analysis Models
6Neural Networks Model
7Prototype Models
8Regression Models
9Support Vector Models

Model Predictions

Once you have computed model outputs, the developed model is saved on the worksheet which you can use to make predictions later. You will need to click the Make Predictions button on the main menu bar to make predictions. A sample screenshot of the main menu bar is shown below.
Analytics Menubar Note that the Make Predictions button is only displayed for certain Analytics worksheets and not all sheets. If you don't see the Make Predictions button on the menu bar, you must click the Refresh button to update the menu bar. This dialog box will help you make predictions for any new data or existing training data.
1
Refresh Button: Click on this button if you don't see the Make Predictions button on the menu bar.
2
Make Predictions: This is an optional button to use if you want to make or update the predictions based on the current model that has been fit to the data.

Setup

If you click on the Setup button within the Make Predictions dialog box, the predictions dialog setup box is displayed. A sample screenshot of the setup tab is shown in this figure. Predictions 1
1
Fitted Model: The model that has been fit for the training data and stored on the worksheet is displayed here. If no model exists or the model was developed for a different data set, you will first need to fit a model before using the predictions module. The fitted model displays the following information:
OptionDescription
ModelThe name of the analytics model that was fit on the worksheet.
NameThe descriptive name of the fit analytics model.
ResponseThe name of the response variable and the type of output data (Binary, Classification, or Regression).
FactorsThe names of the factor variables used in the model development.
AccuracyThe level of accuracy obtained for this model for the test data set.
RSoftwareThe version number of the R software used for the model development.
DateThe date and time that model was last updated.
2
Setup: Displays the setup dialog box. The setup dialog box only contains the destains of the fitted model found on the worksheet and the number of rows of predicted data found on the worksheet.
3
Data: Displays the data dialog box. The data dialog box can be used to upload new prediction data onto the model prediction area of the worksheet. Note that you can import the data from tables or directly type in the input values on the worksheet.
4
Verify: The verify dialog box can check for errors in the prediction model inputs.
5
Predict Data: Shows the number of rows of prediction data defined on the worksheet. Note that only completed rows of data (no missing values) are used to make predictions. If you have missing or unknown input values, use the Pre-Processing tool to estimate the best possible model inputs before using the Prediction module.
6
Help File: Opens the help file for this analysis tool.
7
Cancel Button: This closes the dialog box without making any changes to the settings.
8
OK Button: This tries to run the RScript program to update the predictions for any inputs defined on the worksheet and update the prediction results.

Data

The input data for making predictions is displayed if you click on the Data button within the Make Predictions dialog box. A sample screenshot of the data tab is shown in this figure. Predictions 2
1
Required Columns: The column names used to develop the prediction model are displayed here. You must enter the input data for each column on the worksheet to make new predictions.
2
Available Data: Here, you can see the list of data available in the current workbook. If necessary, you can use the search bar to filter the list of displayed tables and columns. If your input data columns are in any tables, you can drag and drop them to the right. Note that you do not need to import data from tables; you can directly type in the input values on the worksheet.
3
Factor Variables: It shows the list of columns that have been selected to import data. The columns should be specified in the same order as those used for the original model development. For example, if the first column was cyl, the data column containing cyl information should be selected first. For this example, seven columns of input data are required for analysis.
4
Import Data: Once all the columns have been specified, click the Import Data button to import the data from the tables onto the worksheet to make predictions. Note that importing prediction input data from tables is not required, and you can directly copy and paste the required data into the worksheet model predictions area.

Verify

If you click the Verify button, the software will perform some checks on the data you entered. A sample screenshot of the dialog box is shown in the figure below. Verify
1
Item: The left-hand side shows the major tabs and the items checked within each section
2
Status: The right-hand side shows the status of the checks.
3
Overall Status: The overall status of all the checks for the given analysis is shown here. The overall status check shows a green thumps-up sign if everything is okay and a red thumps-down sign if any checks have not passed. Note that you cannot proceed with generating analysis results for some analyses if the overall status is not okay.
If the verification is unsuccessful, then you will not be able to compute the prediction outputs.

OK Button

If there are no verification errors, click the OK button to run the model and update the predictions. The results of the predictions are updated on the worksheet. If you click the Compute Outputs button, the software will reestimate the model by fitting it to the train/test data set and then update the prediction results. A sample screenshot of the prediction result is shown below. Prediction Outputs
1
Heading: Displays the names of the columns for each of the input factors and the output response in the last column.
2
Values: Displays the input values for each factor and the corresponding output response in the last column.

Notes

Here are a few pointers regarding this analysis:
  • This analysis requires that the R software needs to be installed on your computer. Further, you will need to provide a link to the RScript executable file under Sigma Magic Options so that the software can use the R software to generate analysis results.

Examples

The following examples are in the Examples folder.
  • For the data given in the file, perform the bagging analysis for predicting the gear based on other variable inputs. (Bagging Analysis 1.xlsx)



© Rapid Sigma Solutions LLP. All rights reserved.