conditional probability distribution over binary valueΒ π‘Œ is commonly defined by:

  • 𝐏(π‘Œ=1) = π‘ π‘–π‘”π‘šπ‘œπ‘–π‘‘(𝑓(𝑋1, …, 𝑋𝑛))
  • 𝐏(π‘Œ=1) =Β π‘ π‘–π‘”π‘šπ‘œπ‘–π‘‘(𝑀0 + 𝑀1𝑋1Β + … + 𝑀𝑛𝑋𝑛))

where:

  • π‘ π‘–π‘”π‘šπ‘œπ‘–π‘‘(𝒛) = 1Β / (1 + 𝑒-𝒛) #Β is theΒ sigmoid function

similar to: SigmoidΒ Activation Functions (AF)