Machine Learning and Emotions Detection
ML in product development has been exponentially growing in the past few years. ML models are mostly trained on user behaviour, but what about the emotional response to such actions?
A user is clicking on a banner saying "People like you order a bigger size of this item"
Performs a purchase
but feels frustrated because he/she realized that he/she needs a bigger size to be able to wear it.
The model is trained based on successful purchasing actions, but does not detect the frustration feeling.
Is it good to implement this feature?
Will this affect future behaviour in a negative way?
Is there a way to anticipate this i.e. detect and incorporate these emotions in ML models to only show the feature in the future if emotional response is positive?
I am curious to know if anyone with deeper ML knowledge is willing to discuss potential methodologies, experiences and best practices in dealing with a situation like this.
Thank you for this suggestion! Please post a new idea for ProductCamp Berlin 2019!