The Predictive Forecasting Wizard guides you through the simple, three-step process of bringing predictive forecasting technology to your custom work environment. Start by choosing the Time Series Wizard from the Start Page. This will open the Time Series Prediction Overview.
Prediction Overview
The Overview shows the cubes equipped with predictive forecasting. You can also see their database and information relevant to their forecast.
The fields of the Prediction Overview are described below:
| Field | Description |
| Select Database | A dropdown of all databases stored in your Jedox instance. The database is your native work environment populated with custom data. We also provide demo data to allow you to familiarize yourself with driver analysis functionality outside of your own work environment. |
| CUBE (DB) | You may choose any cube for predictive planning purposes, so long as it has at least three dimensions that AIssisted™ Time Series Prediction can utilize as the Time, Version, and Measure dimension. The database is your native work environment populated with custom data. We also provide demo data to allow you to familiarize yourself with predictive forecasting functionality outside of your own work environment. |
| SCENARIO | Shows the saved scenarios for each cube. This can be used to create different time series prediction setups for the same cube. |
| ALGORITHM | AIssisted™ Time Series Prediction uses a number of different algorithms. Each calculates predictive output for your actual data in a unique way. The Best runs through every algorithm possible and chooses the predictions from the algorithm with the best training accuracy. Alternatively, you can prescribe a specific algorithm (Liner Model, Holt Winter, etc.). |
| LAST STARTED | Shows the last time each cube's forecast was started. |
| PREDICTIVE range | Shows the start date, end date, and duration (in months) of the time period undergoing the predictive forecast. |
| SHOW (button) | Shows a preview of the populated data and compares the actual data with your predicted data. It also provides you with an accuracy percentage. |
| START (button) | Runs predictive forecasting for your cube. |
| COPY (button) | Allows you to copy an existing scenario by selecting it from the list of the "Copy Scenario" dialog, or add a new scenario, which will then appear in the "Scenario" list for selection. Click the Save Changes button to confirm the task. |
| EDIT (button) | Allows you to adjust the input data of the cube (e.g. change the predictive range, use a different algorithm, or chose different dimensions/elements). |
| REMOVE (button) | Removes the AIssisted™ predictive scenario and/or prediction values for the cube. Removing only the scenario will allow you to keep prediction values, while removing values will keep the scenario intact. You may also remove both. Click Close in the window once you are finished. |
| SCHEDULE (icon) | Shows a clock icon if the scenario has the scheduled option checked. Click on EDIT to change this option or go to Jedox’s Scheduler section to edit the frequency of the analysis. |
| STATUS (button) | Displays the status of the Integrator job after you start it. To see eventual changes, click the refresh icon in the "Status" column. Once finished, it will show a different icon with the results. Click on the icon for more information. |
| New Setup (button) | Adds AIssisted™ predictive scenario for a cube. |
| Manual Settings (button) | Used to change prediction setups and scenarios without entering the wizard itself. Read more about it in Manual Settings for AIssisted™ Planning. |
Click New Setup to make a predictive forecast. The wizard will guide you through the necessary steps.
Step 1: select and validate cube
The selected database is shown along with a dropdown menu of scenarios and a list of available cubes stored within the database. It also indicates whether each selected cube is validated for forecasting, i.e. whether it has the necessary parts for predictive forecasting to function. A validated cube must have a Time dimension, a Version dimension, and a Measure dimension.
The selected database is shown along with a dropdown menu of scenarios and a list of available cubes stored within the database.
The "Cubes" list allows you to choose the particular cube on which you would like to perform time series prediction. The selection will indicate with a green or red message box whether the selected cube is validated for the analysis, i.e., whether it has the required dimensions. A validated cube must have a Time dimension, a Version dimension, and a Measure dimension. Click the Advanced Setting button to access the dimension type assignments.
The "Scenario" Combobox gives you the option to save more than one scenario per cube. Select your scenario from the dropdown menu or click the + to open the "Create New Scenario" dialog. There you can add a new scenario, which will then appear in the "Scenario" list for selection. Click the Save Changes button to confirm the task.
Once you have selected a validated cube, click Confirm Selection and then Next.
Step 2: select Prediction Source
You have selected your cube. Now you must narrow your selection to a specific data slice.
The essential components of this slice are the Time, Version, and Measure dimensions, located in the upper portion of the wizard. The Other Dimensions, found in the lower portion, will further fine-tune your prediction. More exact specifications make your data slice smaller and may make your predictions more accurate.
The Set defaults button dynamically selects the first and last date elements as well as the topmost elements for additional dimensions. You may, of course, also select your source material manually. Simply select the start and end Time dimensions from the Comboboxes, being sure to choose the time level at which your data resides. For example, if your data is daily, choose a daily date, e.g. 2021-01-01. If you would like to schedule your predictions to run on a certain interval with the newest data, click the checkbox “Use last month as end date”. Then click the adjustment icons (
) to select the rest of your source material, which can include multiple elements per dimension, with the exception of the Time dimensions. Note: when setting your prediction source, more specific element selections may result in more accurate forecasts.
The Mode setting for other dimensions has two options: onlyNodes and onlyBases. This sets whether the algorithms will use data at the consolidated (node) level or at the base level. If the mode is set to onlyNodes, the prediction results will be splashed down to the base level using the same distribution as the last period of actual data.
Because using data at the base level may result in extreme quantities of data used in predictions and long wait times for prediction results, the wizard limits the number of dimensions that can be set by onlyBases. If more than one dimension is chosen, an error will appear when trying to move on.
If you are sure you would like to use base level data for more than one dimension, this can be set manually in the Manual Settings report. However, it is recommended that first the predictions be run with the onlyNodes mode.
The Source View button allows you to view the source data before going through the prediction process. After entering your parameters, click on this button, and a preview window will pop up, showing how the source data looks. In case of gaps or extreme values, it can be helpful to use the Outlier detection on step 3, or the AIssisted™ Data Preparation Wizard.
If you are satisfied with the data, just close the "Preview Source" window. When you are finished selecting your prediction source, click Next.
Step 3: select Prediction Properties
Now you must select the Prediction Properties, namely the prediction interval, duration and type.
You can choose to have monthly data or daily data (only select if daily data is available in the cube) for your prediction intervals and a predictive duration of 1-24 months. There are various prediction types, each with different algorithms. The Classic prediction type runs statistical methods, while the Innovate prediction type implements machine learning. The Best Selection prediction type runs all algorithms and selects the one that is the most accurate.
Select the checkboxes to apply either of the predictive management tools. Outlier detection controls for inconsistencies, thereby providing more accurate data. The Accuracy cube stores the accuracy of the prediction for the chosen algorithm. When Best Selection mode is selected, the accuracy values for all algorithms are stored in the Accuracy cube and the algorithm with the highest accuracy is chosen. Schedule Prediction automates a new prediction for a set interval. It will refresh your prediction along with incoming data at a set frequency. Using the checkbox in Step 2, the scheduled job will always use the latest data. The default frequency is once per day, but you can customize the length of your prediction intervals in the Scheduler.
Once you have selected your Prediction Properties and chosen the predictive management you wish to incorporate, click Next.
Summary Step:
Here you see a summary of your scenario and a small preview of your source data. You may now save and execute the Time Series Prediction or simply save your prediction settings. Either way, your forecast for this data set will now appear in the Prediction Overview. Here it can be viewed, modified, and executed at any future point.
Once executed, the data automatically becomes available in the cube and can be implemented in reports and templates.
Time Series Prediction Wizard with R services
Time Series Prediction Wizard setups all use Python services by default since December 2023 with version 23.4.0. The R services will continue to run, however the Integrator jobs that connect to those services in the Time Series Prediction Wizard will no longer be maintained after version 25.1.2.
No action is required with setups created with Wizard version 23.4.0 or later. You can double check whether you are using R or Python in the Manual Settings for AIssisted™ Planning. See screenshot below:
For setups created before then, to be sure you are happy with the results, create a copy of the scenario so you can compare results. After it is copied, go to Manual Settings, AIssisted sheet, and change the TSPodType to P. (Be sure the database, cube, and scenario drop-downs are correct.)
Then open the newly created Wizard scenario and change the prediction target version (to ensure you don't write over the R results). Click through to the end and execute.
Compare results. Bounds will likely be narrower because the calculation is different. What we are pulling from the service is called a confidence interval, which is highly dependent on the data and the algorithm used. It is so much narrower because it is giving its most confident calculations. With R, these bounds were more or less calculated in a static way (20% up or down with an additionally smaller percentage added to this calculation, depending on how far out it was into the future).
There is also another calculation the Python service does that ends up returning a larger interval that would be more similar to the R service values we are used to seeing, which is called the prediction interval. You can change which calculation the Wizard takes by adjusting the TSBounds calculation value in the Manual Settings report on the AIssisted sheet. You can also adjust the interval width easily from the Manual Settings. At this point, the default is set at the recommended 0.2 interval.
When you are happy with your results, delete the old scenario with the R services and schedule the new scenario (if applicable).
Updated November 18, 2024