Perspective on Automation of Statistical Modeling Process for Battery Lifetime Prediction

Automating the process of statistical modeling would be substantially beneficial for the fields that rely heavily on expert statisticians. In this paper, a perspective is briefly provided on how to automate statistical modeling processes for the prediction of battery lifetime, especially on automating the step of feature engineering, to alleviate the considerable time and human intervention required by the current data-driven lifetime prediction methods. An example related to overpotential concept for early prediction is discussed to show that teach machines concepts the same way as humans are taught may be a promising way to enable the machine to gain anthropomorphic intelligence that is beneficial for automated feature engineering.