Measuring Innovation Patterns in Iran and Neighboring Countries: A Time Series Similarity Approach Using STL and Dynamic Time Warping

Abstract

Innovation is becoming ever more pivotal to national development strategies but measuring and comparing innovation performance across nations is still a methodological challenges. This research devises a new time-series similarity method that integrates Seasonal-Trend decomposition (STL) with Fast Dynamic Time Warping (DTW) to examine Irans innovation trends by comparison with its regional peers. Owing to data availability constraints of Global Innovation Index data , research and development spending as a proportion of GDP is used as a proxy with its limitations clearly noted. Based on World Bank indicators and an Autoencoder based imputation technique for missing values, the research compares cross-country similarities and determines theme domains best aligned with Irans innovation path. Findings indicate that poverty and health metrics manifest the strongest statistical similarity with R and D spending in Iran, while Saudi Arabia, Oman, and Kuwait show the most similar cross country proximity. Implications are that Iranian innovation is more intrinsically connected with social development dynamics rather than conventional economic or infrastructure drivers, with region-specific implications for STI policy.