Abstract:
Climate and global warming play a crucial role in the lives of living organisms on Earth.
Temperature, varying in each region, is a vital aspect in climate observation. This study analyzed
temperature fluctuations in Pekanbaru from 2016 to 2022 using fractal analysis and
Backpropagation artificial neural networks. The research findings revealed that temperature
prediction with Backpropagation artificial neural networks was quite accurate. However, errors
during testing or validation could impact the comparison with the target values. Fractal analysis
indicated a persistence tendency in temperature fluctuations in Pekanbaru, with a Hurst exponent
of 0.7993 and a fractal dimension of 1.2007. Nevertheless, temperature fluctuations were also
influenced by other factors, leading to varying levels of stability over certain periods. Thus,
temperature in Pekanbaru can be considered a complex system with diverse fluctuation patterns
and varying levels of complexity.
