Trend analysis and Applicability of Fuzzy Time Series Models to Night-time Luminosity Data of Uttarakhand state, India

Abstract

Objectives: Application of trend detection and fuzzy time series models to predict the growth of Uttarakhand state using night-Time Lights (NTL) dataset (Visible Infrared Imaging Radiometer Suite (VIIRS) night-time lights). Methods: In this work, we propose a methodology to analyze trends and estimate different fuzzy time series models for prediction in the context of forecasting the economic growth of a low-earning income Uttarakhand state. Mann- Kendall tests are used for performing trend analysis of Uttarakhand (India) using night-time luminosity data sets. Forecasting methods based on fuzzy time series models have been presented to deal with uncertainties induced by ambiguities, vagueness, and some other non-probabilistic factors, and have widely been employed in many problems domains. The simulation is done on the Jupyter Notebook platform using night-time luminosity datasets. Findings: We conclude that Uttarakhand and Himachal Pradesh show positive growth in trends for their night-time lights monthly average intensity with magnitude of 0.002182 and 0.002558 respectively. Night time luminosity database can be used as a proxy indicator for the economic growth of the Uttarakhand state as the correlation between NTL time series data of Uttarakhand and its Gross Development Product (GDP) is 0.795641021. Fuzzy time series models can also be used to predict the growth of the Uttarakhand state and the High Order Fuzzy Time Series model gives the accuracy metrics of root mean square (RMSE) as 0.11. Novelty: No previous research work is performed on trend analysis and prediction of growth of a Uttarakhand state using night-time luminosity data sets. Keywords: Fuzzy Systems; Fuzzy Time Series; Prediction; Night time luminosity; Satellite image