Machine Learning-Based Prediction of Circular Economy Implementation Effectiveness in Education for Sustainable Development (ESD) Programs: Empirical Evidence from Indonesia

Imagine the world as a giant, messy workshop where we constantly build things, use them, and then throw them away. For a long time, we've operated like a one-way street: take resources, make products, and dump the leftovers. But scientists and educators are now trying to turn this into a giant, looping carousel called the Circular Economy. Instead of trash, everything gets recycled, reused, or turned into something new, keeping resources spinning in a happy circle rather than disappearing into a landfill. To make this happen, we need to teach people how to think in loops. This is where Education for Sustainable Development (ESD) comes in. Think of ESD as the school curriculum that trains our brains to see the world as a connected system where nothing is wasted. But here's the tricky part: just because a school says it's teaching these ideas doesn't mean it's actually working. How do we know if the "circular" lessons are sticking? That's the big question. Instead of just asking teachers, "Did you do a good job?", a new study decided to use a super-smart digital detective called Machine Learning to peek behind the curtain and see what really makes these programs succeed or fail.

---

The Digital Detective and the School Loop

A team of researchers from Indonesia decided to play detective with a very specific mystery: How well are schools in Jakarta and Banten actually implementing these "circular economy" ideas? They didn't just guess; they gathered a massive team of 505 people, including teachers, students, school bosses, and government policymakers, to fill out a survey. They wanted to see if the "circular" lessons were actually effective or if they were just buzzwords on a chalkboard.

To solve this, they built a Machine Learning model. If you imagine the data from the survey as a giant pile of puzzle pieces, traditional math is like trying to fit them together by looking at one piece at a time. Machine Learning, however, is like a robot that can look at the whole pile at once, finding hidden patterns and connections that a human eye might miss. The researchers fed their puzzle pieces into three different types of robot detectives: a Decision Tree (which asks a series of yes-or-no questions), a Random Forest (which is like a whole forest of trees voting on the answer), and a Support Vector Machine (which tries to draw the best possible line to separate "good" programs from "bad" ones).

What They Found: It's a Team Effort, Not a Solo Act

The results were a bit surprising. First, the "effectiveness" of these programs wasn't a wild rollercoaster of highs and lows. Most of the time, the programs were sitting right in the middle—moderate. It wasn't a disaster, but it wasn't a perfect utopia either. The scores hovered around a middle ground, suggesting that while schools are trying, they haven't quite mastered the art of the circular loop yet.

The researchers then asked the robots: "What makes the difference?" They expected maybe one big factor, like having a fancy new computer lab or a super-strict policy, to be the magic wand. But the robots said, "Nope, it's a team effort."

The models found that no single factor ruled the roost. Instead, a bunch of things worked together:

• Teacher Competence: How good the teachers are at explaining these concepts.
• Institutional Support: Whether the school administration actually backs the program.
• Curriculum Integration: How well the circular ideas are baked into the actual lessons.
• Student Engagement: Whether the kids are actually paying attention and participating.
• Policy Alignment: Whether the government rules match what the school is doing.
• Technological Support: Having the right tools and tech.

The study suggests that if you only fix the technology but ignore the teachers, or if you have great teachers but no school support, the program won't work well. It's like trying to bake a cake with only eggs and no flour; you need all the ingredients to get a good result.

The Robot's Verdict

When it came to predicting how well a program would do, the results showed that the different robot detectives performed at relatively comparable levels. While the Random Forest model did show slightly more stable results and a small edge in accuracy (predicting effectiveness about 85% of the time in tests), the paper notes that no single model demonstrated a clearly dominant performance across all metrics. The Decision Tree and Support Vector Machine were also quite capable, suggesting that the underlying data is complex and that different tools can offer similar insights rather than one "star" detective solving the case alone.

Interestingly, when the robots looked at which factor was the most important, they pointed to Teacher Competence as the heavyweight champion, followed closely by Institutional Support and Curriculum Integration. This suggests that having a skilled teacher who knows how to teach these complex ideas is slightly more critical than having the perfect policy or the newest gadgets. However, the difference wasn't huge; all the factors played a role.

The Bottom Line

This study didn't find a magic button that instantly fixes sustainability education. Instead, it suggests that making circular economy programs work is a bit like conducting an orchestra. You need the teachers (the violinists), the school leaders (the conductors), the curriculum (the sheet music), and the students (the audience) all playing in sync. If one section is out of tune, the whole song suffers.

The researchers used these digital tools to show us that while we are making progress, there is still a lot of work to do. The "moderate" scores mean we are on the right path, but we need to keep tuning our instruments. By using these smart computer models, educators and policymakers can now see exactly where to focus their energy—likely on training teachers and supporting schools—rather than guessing. It's a reminder that saving the planet isn't just about big ideas; it's about the small, interconnected details of how we teach and learn.