ABSTRACT
BACKGROUND: Emotions and mood are important for overall well-being. Therefore, the search for continuous, effortless emotion prediction methods is an important field of study. Mobile sensing provides a promising tool and can capture one of the most telling signs of emotion: language. OBJECTIVE: The aim of this study is to examine the separate and combined predictive value of mobile-sensed language data sources for detecting both momentary emotional experience as well as global individual differences in emotional traits and depression. METHODS: In a 2-week experience sampling method study, we collected self-reported emotion ratings and voice recordings 10 times a day, continuous keyboard activity, and trait depression severity. We correlated state and trait emotions and depression and language, distinguishing between speech content (spoken words), speech form (voice acoustics), writing content (written words), and writing form (typing dynamics). We also investigated how well these features predicted state and trait emotions using cross-validation to select features and a hold-out set for validation. RESULTS: Overall, the reported emotions and mobile-sensed language demonstrated weak correlations. The most significant correlations were found between speech content and state emotions and between speech form and state emotions, ranging up to 0.25. Speech content provided the best predictions for state emotions. None of the trait emotion-language correlations remained significant after correction. Among the emotions studied, valence and happiness displayed the most significant correlations and the highest predictive performance. CONCLUSIONS: Although using mobile-sensed language as an emotion marker shows some promise, correlations and predictive R2 values are low.
ABSTRACT
New robust correlation models for ozonation, based on UVA254 and fluorescence surrogate parameters and developed considering kinetic information, have been applied at pilot-scale. This model framework is validated with the aim for operators to control the ozone dose for the removal of trace organic contaminants (TrOCs) in effluents from full-scale municipal wastewater treatment plants. The inflected correlation model between ΔTrOCs and the surrogates predicts the removal of TrOCs (based on statistical evidence) solely using the 2nd order reaction rate constant with ozone (kO3) and in a more adequate manner than similar single correlation models. This allows the use of this new model for current and future TrOCs under investigation which is highly interesting when imposed discharge limits might include more and other TrOCs in future. The use of UVA254 might be preferable at the current timing for online monitoring of TrOC abatement as the model showed a good predictive power (based on statistical evidence and visual confirmation). Reliable online sensors are more widespread (and commercially) available compared to fluorescence sensors which are still under development, with the exception of a few examples. Nevertheless, the data processing of the fluorescence signals, isolating the different intensities associated with moieties reacting similarly to ozone might even increase the predictive power, given the lower degree of interference (i.e. less scattering).
