Process cubes, introduced by Wil van der Aalst, are a technique used in multi-dimensional process mining, a subfield of process mining that focuses on analyzing processes from multiple perspectives or dimensions. They provide a holistic view of a process by aggregating and visualizing process data along various dimensions, such as time, resources, and data attributes. Here's how process cubes differ from traditional process mining approaches:

1. **Holistic View**: Traditional process mining approaches often focus on a single perspective, such as the control flow (what activities are executed and in what order), the organizational perspective (who performs the activities), or the case perspective (the data associated with individual cases). Process cubes, on the other hand, combine and visualize data from multiple dimensions, providing a comprehensive view of the process.

2. **Aggregation**: Process cubes aggregate process data along the chosen dimensions. This aggregation helps to identify patterns and trends that might not be visible at the individual case level. In traditional process mining, aggregation is often done post-analysis, while process cubes integrate aggregation into the mining and visualization process.

3. **Visualization**: Process cubes use a 3D cube metaphor to visualize the aggregated process data. This visualization allows analysts to easily explore and understand the process from different angles. Traditional process mining approaches typically use 2D visualizations, such as Petri nets, BPMN diagrams, or parallel coordinates, which may not provide the same level of multi-dimensional insight.

4. **Dynamic Analysis**: Process cubes can be used to analyze dynamic processes, where the behavior can change over time. By including time as a dimension, process cubes can show how the process evolves and changes. Traditional process mining approaches often focus on static processes or use complex techniques to analyze dynamics.

5. **Flexibility**: Process cubes allow analysts to choose the dimensions they want to analyze, making them highly flexible. This flexibility contrasts with traditional process mining approaches, which often focus on a predefined set of perspectives or dimensions.

Here's a simple example to illustrate the difference:

- **Traditional Process Mining (e.g., using a Petri net)**: You might analyze the control flow of a process and see that 'Activity A' is always followed by 'Activity B'. This tells you about the order of activities.

- **Process Cube**: You might create a process cube with dimensions 'Time', 'Resource', and 'Data Attribute'. By analyzing this cube, you might discover that 'Activity A' is always followed by 'Activity B', but only when performed by 'Resource X' and when the 'Data Attribute' is 'High'. This provides a more nuanced understanding of the process, considering multiple dimensions.

In summary, process cubes in multi-dimensional process mining provide a more holistic, aggregated, and flexible view of processes compared to traditional process mining approaches. They enable analysts to explore and understand processes from multiple dimensions simultaneously.