Using Speech Technology for Quantifying Behavioral Characteristics in Peer-Led Team Learning Sessions.

•Established CRSS-PLTL corpus (Peer-Led Team Learning) and performed exploratory data analysis.•Stacked Denoising Autoencoder-based bottleneck features + Informed HMM-based diarization system.•Behavioral Speech Processing for extracting characteristics such as participation, dominance, curiosity (in terms of question inflection), emphasis, engagement.•Stacked spectral features were used to train a Deep Neural Network for estimating the fundamental frequency.•Fundamental frequency-based method for question inflection detection.