Free cookie consent management tool by TermsFeed Models [Process Mining v5.8]
Models
This content applies solely to Process Mining, which must be purchased separately from the Appian base platform.

Models

Models are a visual representation of your current or ideal processes. Models help to identify common bottlenecks, deviations, and other impediments to successful and efficient completions. One of the primary forms of analysis in Process Mining is conformance checking, where differences between the discovered and target models are highlighted.

There are two model types:

  • Discovered model: The visualization of your existing processes derived from event logs and optionally, case attribute files.
  • Target model: The visualization of the improved, but not yet realized, process.

To view the models for your processes, click the Model icon in the navigation menu. You can switch between the discovered and target models by clicking their respective tabs.

pm-model-tab

Discovered model

When you first sign in to Process Mining and select a log, Appian displays your discovered model. This is a visualization of your process based on the data in your event log.

The activities in your process are represented as nodes in the model. The connections between activities are represented as connectors.

By default, the size of activity nodes and the connectors between them represent the frequency the activity occurred. This helps you quickly see how often activities occur in relation to one another. For example, if one activity occurs 200 times and another occurs 100 times, the radius of the activity node that occurs 200 times is twice as large as the radius of the activity node that occurs 100 times.

The color of activity nodes represents the average duration of the activity based on minimum and maximum values in the event log. For example, connectors with an average duration closer to the minimum are blue, and connectors with an average duration closer to the maximum are red.

Tip:  You must identify start and end time stamps in your event log to see activity duration metrics in your model.

To change how your discovered model is visualized, you can adjust the visual settings.

Adjust visual settings

There are many options to customize the appearance of your models. Adjust the model settings to gain new insights into your processes.

To change your discovered model settings:

  1. Click the Discovered Model tab.
  2. Click the Model settings icon.
  3. Configure the size and color metrics.
    • The Size metric determines the property that changes activity node sizes. Options include Frequency, Duration, and Case Count.
      • For Duration, select whether to display the Minimum, Maximum, Average, Median, Total, or Standard deviation.
    • The Color metric determines the property that changes the color of activity paths. Options include Frequency, Duration, Case Count, Throughput, and Deviation to target duration.
      • For Duration, select whether to display the Minimum, Maximum, Average, Median, Total, or Standard deviation.
      • For Deviation to target duration, select whether to display the Minimum, Maximum, Average, Median, or Total.
      • For Throughput, select the time interval (second, minute, hour, day, week) and time span using the start and end times.
  4. Click OK.

pm-model-settings

If your models are particularly complex, you may wish to hone in on a specific activity. Nodes represent activities in Process Mining models. Select an activity from the Find Activity drop-down, and your view will zoom into the respective node.

pm-find-activity

You can also simplify the discovered model with the complexity slider that is located in the bottom-left of the window. Drag the slider towards Simple and away from Complex, and the model automatically adjusts.

When you simplify the model, less frequent connectors between activities are removed. Reducing the complexity doesn't change or remove the activities themselves. This is especially useful if you want to visualize several variants of a model. By default, the process displays all of its follower relationships. In other words, it displays the model in its full complexity.

pm-simplify-model

When you click on an activity node, Process Mining displays more detailed information, such as:

  • Number of activities over the displayed cases
  • Minimal duration
  • Maximum duration
  • Average duration
  • Median duration
  • Standard deviation

pm-activity-info

You can also select from a few additional options:

  • Filter activity: Only display cases that include the selected activity. You can select multiple activities that must be included in the cases.
  • Filter activity duration: Filter based on minimum and maximum duration for the activity. This option is only available if your event log contains start and end time stamps.
  • Hide this activity: Hide this activity from the model. You can select multiple activities to hide.

Add target durations

Target durations help you track and visualize how the durations of activities in your actual process compare with a goal, industry standard, or service-level agreement (SLA). For example, you might want to review the average response times in an incident management process to verify that you're following your SLA.

You'll need to create a CSV file that defines the target durations for specific activities or between activities. A Path Duration File defines target durations between activities. An Activity Duration file defines target durations for specific activities.

To create a Path Duration file, create a CSV that contains the following columns:

  • From: The name of the activity that starts the path.
  • To: The name of the activity that ends the path.
  • Target: The target duration between From and To activities. Valid units include milliseconds, seconds, minutes, hours, days, months, or years.

To create an Activity Duration file, create a CSV that contains the following columns:

  • Activity: The name of the activity.
  • Target: The target duration. Valid units include milliseconds, seconds, minutes, hours, days, months, or years.

Once you've created the file or files that define your target durations, you can add them to the discovered model.

To add target durations:

  1. Go to DISCOVERED MODEL.
  2. In the header bar, click the event log name.
  3. Select Add target durations.
  4. Click the paperclip icon next to the Path Duration file or Activity Duration file.
  5. Open the relevant file.

    Tip:  You can add Path Duration and Activity Duration at the same time.

  6. Use the dropdowns for each column header to select the data the column contains. Configure Path Duration
  7. Select the unit of time for the target duration in Duration unit.
  8. If you've added both Path Duration and Activity Duration files, click NEXT and repeat steps 6-7.
  9. Click FINISH.

Now, new visual settings are applied to the discovered model to visualize the target durations. The colors of the activity nodes and activity connectors change depending on the target durations you set.

  • White: No value.
  • Dark Blue: Lower duration than the target.
  • Light Blue: Meets the target duration.
  • Red: Higher duration than the target.

From here, you can adjust the visual settings to track against minumum, maximum, average, median, or total durations.

Animate the model

Process Mining can animate the discovered model process. The animation shows how cases run through specific process variants.

To animate the discovered model:

  1. Click the Discovered Model tab.
  2. Click the Show animation icon to display animation controls.
  3. Click the + or - Animation speed buttons to speed up or slow down the animation.
    1. Note: When the animation speed is set to 1, the animation is scaled to display all log content in one minute.
  4. Click Play animation. To stop the animation, click Stop animation.

pm-model-animation.png

Target model

The target model represents the ideal reference model for a process. You can upload the target model as a BPMN file or create it directly in Process Mining.

Note:  To compare the duration of individual activities or the duration between activities with a target duration, add target durations to the discovered model.

pm-target-model.png

To see the target model, click the Target Model tab in Process Mining. If you haven't connected a target model yet, the first time you view the Target Model tab, Process Mining will prompt you to select an existing model, or:

Whether you choose an existing model, upload a new model, or create a new model, the target model activities must match the events in the discovered model. Process Mining matches activities by name, so make sure that the names of your activities in the discovered model and events in the target model match. If you try to connect a target model with activities that don't align with the discovered model, Process Mining displays the misaligned activities.

If you already connected a target model and you want to change it, go to Model Management.

Upload a target model

To upload a new target model:

  1. Click the Target Model tab.
  2. Click Upload New Model.
  3. Choose the BPMN file that represents your target process.
  4. Click Open.
  5. Verify that the activities align with the discovered model and then click Save Alignment.

If the activities do not align, no data will be associated with the misaligned activities.

pm-new-target-model.png

Create a target model

You can create or edit Business Process Model and Notation (BPMN) models right in the system using built-in bpmn.io functionality. Additionally, Process Mining can generate a model based on the discovered model.

To create a new target model:

  1. Click the Target Model tab.
  2. Click Create New Model.
  3. Select Yes or No, depending on whether you want to start with a model inferred from the discovered model or start from scratch.
  4. Modify the model as needed using standard BPMN modeling tools.
  5. Click SAVE MODEL.
  6. Enter a name for target model. BPMN Model displays by default.
  7. Click SAVE MODEL to upload the model to Process Mining and associate it with the current event log and discovered model.

Create target model

The icons in the bottom-left of the window allow you to start a new model from scratch, return to the discovered model, upload a model, download the target model BPMN file, or save the target model to Process Mining.

Adjust visual settings

By default, the target model darkens activity nodes based on activity duration and thickens the connectors between them based on the frequency they occur in your discovered process. Connectors that are thinner than others indicate that path occurs less frequently.

The color of activity nodes can range from:

  • White: No duration.
  • Blue: Low duration.
  • Red: High duration.

For a more precise view, click an activity or connector to see the associated numeric attributes that can impact the visual representation.

To change your target model settings:

  1. Click the Target Model tab.
  2. Click the Model settings icon.
  3. Configure the path thickness and activity color setting. Options include Frequency, Minimum Duration, Maximum Duration, Average Duration, and Standard deviation.

Adjust variant groups

In addition to the size and color of activity nodes and connectors, you can also adjust the number of variant groups the model displays.

  • Variant groups are variants that are grouped according to their frequency. In other words, all variants with the same frequency are considered a variant group. Variant group 1 always represents the most frequent variants. Variant group n represents the least frequent variants.
  • Variants are groups of cases that have the same order and sequence of activities.

To adjust model variant groups, drag the Min and Max sliders in the Number of Variant Groups section to the number you want to display. If all variant groups are visualized by the model, the complete event log displays in the discovered model.

If several process variant groups are represented in the target model, process deviations are also visualized:

  • Dotted red line: Alternate process paths and sequences
  • Red plus icon: Activities were performed multiple times. For example, because there is a loop in the process.
  • Blue fast-forward icon: Activities were skipped in some variant groups.
  • Orange swap icon: Activities occurred in a different order.

pm-deviation-type.png

Note:  You can filter by single variants on the Variants page.

Manage models

You can add, share, edit, export, and delete target models from the management hub. The TARGET MODELS tab displays all of the target models that you've added to Process Mining or that have been shared with you.

Note:  Discovered models are automatically generated from an event log, while target models are BPMN files you upload or create. Models in this section only relate to target models except for exporting discovered models.

Access target models

To access target models:

  1. Click Management Hub in the sidebar.
  2. Go to TARGET MODELS.

Upload a target model

You can upload target model to Process Mining as BPMN files.

To upload a model:

  1. Click Management Hub in the sidebar.
  2. Go to TARGET MODELS.
  3. Click NEW.
  4. Find and open the BPMN file from your computer.

When you upload a target model, it isn't connected to an actual process. To connect an uploaded target model with a process, connect or update an event log on the LOGS tab.

You can also upload a target model on the Models page directly if one isn't already connected.

Edit a target model

Process Mining supports viewing and editing BPMN files directly in the system.

To edit a model:

  1. Click Management Hub in the sidebar.
  2. Go to TARGET MODELS.
  3. Select a target model.
  4. Click EDIT.
  5. Edit the target model using the standard BPMN tools.
  6. Click SAVE MODEL.
  7. Optionally, change the name of the updated model.
  8. Click Save Model.

Create target model

Tip:  To edit the model directly from the Target Model tab, click Edit Model.

Share a target model

You may want to share a model with your entire organization or with specific users.

To share a model:

  1. Click Management Hub in the sidebar.
  2. Go to TARGET MODELS.
  3. Select a target model.
  4. Click SHARE.
  5. Select the radio button to indicate whether you want to share the model with your entire organization or with individual users.
    • If you selected Share with individual Users, select the checkboxes next to the users in the list.
  6. Click SAVE.

pm-share-model

To stop sharing a model:

  1. Click Management Hub in the sidebar.
  2. Go to TARGET MODELS.
  3. Select a target model.
  4. Click UNSHARE

Export a model

When you export a model from Process Mining, it downloads a file on your computer. Target models export as BPMN files and discovered models export as either BPMN, PDF, or SVG files.

You can either export target models from the management hub or export any type of model from the EXPORT dropdown in the header bar.

Export a model from the management hub

To export a target model:

  1. Click Management Hub in the sidebar.
  2. Go to TARGET MODELS.
  3. Select a target model.
  4. Click EXPORT.

Note:  When you export a target model and have filter settings applied, the filter settings will also be exported.

Export a model from the EXPORT dropdown

The EXPORT dropdown displays in the header bar whenever an event log is active.

  • The export discovered model option is available when an event log is active.
  • The export target model option is available when the active event log is connected to a target model.

pm-export-menu

To export a target model:

  1. Click the EXPORT dropdown in the header bar.
  2. Select Target model.

To export a discovered model:

  1. Click the EXPORT dropdown in the header bar.
  2. Select Discovered model.
  3. Enter the name you want for the file.
  4. Select the file type you want to export..
  5. Click DOWNLOAD.

pm-export-menu

Delete a target model

To delete a target model:

  1. Click Management Hub in the sidebar.
  2. Go to TARGET MODELS.
  3. Select a target model.
  4. Click DELETE.

If you delete a target model, this action cannot be undone.

Open in Github Built: Thu, Mar 28, 2024 (10:34:14 PM)

Models

FEEDBACK