SHORT SUMMARY
The article discusses the development of a web-based corruption case mapping system using machine learning, specifically an Artificial Neural Network (ANN) with a backpropagation method. The system aims to aid corruption prevention in Indonesia by visualizing corruption cases per region based on news data.
Key Components:
- Data Collection:
- Over 900,000 news articles were gathered through web crawling and scraping from seven major Indonesian news portals.
- Articles were classified into corruption-related or non-corruption-related categories.
- Machine Learning:
- ANN with backpropagation was used for classification, achieving an accuracy of 96.91% with a Sigmoid activation function.
- The input data utilized a “Bag of Words” model, and the ANN model had two hidden layers.
- Visualization:
- The application uses Google Map API to display corruption cases geographically.
- Features include:
- Color-coded severity levels for quick visual interpretation.
- Historical charts showing changes in corruption cases over time.
- Web Application Development:
- Built using the Laravel framework, the web application provides interactive maps and detailed insights.
Key Takeaways
- Significant Impact of Corruption: Corruption has led to a loss of USD 15 billion in Indonesia as of 2016, emphasizing the urgency of prevention efforts.
- Integration of Technology: The study highlights how machine learning and visualization tools can be leveraged for societal challenges like corruption.
- High Accuracy of ANN: The ANN model with a Sigmoid activation function was particularly effective in classifying corruption-related news, demonstrating the potential of machine learning in similar applications.
- Visualization as a Decision-Making Aid:
- The interactive map and historical data provide stakeholders with detailed insights into corruption patterns, facilitating targeted interventions.
- Future Research Directions:
- Expanding the algorithm comparison beyond activation functions to include other machine learning methods.
- Enhancing computational resources for more robust validation methods, such as k-fold validation.
Who Can Benefit from This Research?
-
1. Government and Anti-Corruption Agencies
- Corruption Eradication Commission (KPK): The tool provides real-time, region-specific insights into corruption cases, helping prioritize investigations and allocate resources efficiently.
- Policy Makers: Enables data-driven decisions for crafting anti-corruption strategies and evaluating the effectiveness of existing policies.
2. Law Enforcement and Judicial Bodies
- Police and Prosecutors: Assists in identifying corruption hotspots and trends to build stronger cases.
- Courts: Provides contextual evidence for corruption-related trials.
3. Public Sector Organizations
- Ministries and Local Governments: Helps in tracking corruption trends within specific regions or administrative levels, allowing for targeted reforms.
- Procurement Departments: Insights into corruption in procurement processes can enhance transparency and accountability.
4. Businesses and Corporations
- Private Sector Entities: Businesses can use the tool to assess risks when operating in certain regions and avoid involvement in corrupt practices.
- Whistleblowers: The application facilitates a culture of transparency, indirectly supporting whistleblowing efforts.
5. Researchers and Academics
- Social Science and Public Administration Researchers: The data can serve as a foundation for further studies on corruption dynamics and prevention strategies.
- Data Science and Machine Learning Experts: The research showcases how machine learning can be applied to social issues, encouraging similar innovative applications.
6. Civil Society and NGOs
- Anti-Corruption Organizations: Helps advocacy groups track corruption trends and push for accountability in specific regions.
- Community Groups: Raises awareness of corruption prevalence in their local areas, empowering citizens to demand change.
7. General Public
- Awareness: The tool educates citizens about corruption patterns in their regions, fostering greater demand for transparency and ethical governance.
- Participation: Provides a platform for individuals to report and understand the scale of corruption.
8. Media Outlets
- Journalists can use the insights to investigate and report on corruption-related issues, contributing to public discourse and holding authorities accountable.
9. Technology Developers
- Encourages IT professionals and startups to explore innovative uses of machine learning and visualization tools for societal problems.
