Insights

•θ ∝ h_θ(x)

•Since we cannot vary θ, we can change our input x to reduce the probability and bring it closer to 0 (Acceptance)

•Higher weights mean higher probability (closer to 1 i.e., Denial)

•For a one hot encoded model, we can obtain the values of every feature which contribute more to 0 (Acceptance)

•These are the values we recommend under insights to increase an applicant’s chances of Acceptance