ANALYSIS OF CASHLESS SOCIETY STATUS OF A COUNTRY BY USING THE RANDOM FOREST CLASSIFIER METHOD ON MACHINE LEARNING

• Main Author: Zahrotun Nihayah, Fajris
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/79618
• Institution: Institut Teknologi Bandung

The rapid development of digital financial services has been driven by technological advancements, which have enabled the emergence of innovative digital financial services. Starting from the development of ATMs, debit and credit cards, digital wallets, to the adoption of QR payment systems in Indonesia, which have been standardized under the name QRIS. The emergence of alternative payment methods has created a new habit among society to rely less on cash. This change in payment patterns has led to the development of new knowledge about online transactions and their impact on social welfare in each country. In this study, a process of classifying countries into cashless, almost cashless, and not cashless based on the Cashless Transaction Index (CTI) using the Exponential Moving Average (EMA) algorithm and Random Forest Classifier (RF) with 1225 quantitative parameters such as GDP per capita, electronic transaction volume, the number of digital transaction users, the number of ATM users, and other parameters like digital infrastructure availability and internet penetration in each region. The classification results yielded several cashless countries, such as Canada, New Zealand, Denmark, the United Kingdom, and other developing countries with a CTI value greater than 3.0. On the other hand, countries like Turkmenistan, Congo, and Yemen are considered non-cashless countries with a CTI value of 0. Indonesia itself still maintains a not cashless status despite rapid growth in electronic payment development, hence the government should try to develop a hybrid transaction system.