Data-Driven Prediction and Maturity Assessment of Water Electrolysis Hydrogen Production Technologies Based on LSTM-SVR and TRIZ Theory

Water electrolysis is key to clean hydrogen production, yet its technological evolution remains uncertain. This study applies LSTM-SVR modeling on patent data and TRIZ S-Curve analysis to predict the 15 -year trajectory of ALK, PEM, SOEC, and AEM. Results show LSTM-SVR outperforms single models, accurately capturing trends. AEM and PEM are emerging as mainstream technologies, SOEC remains nascent, and ALK is maturing with declining market share. These findings provide scientific guidance for R&D, policy, and industry strategy.