Hot News Recommendation System from Heterogeneous Websites Based on Bayesian Model

Algorithm 1

Hot news recommendation algorithm.

Hot news recommendation algorithm based on titles

begin

T ← ally of titles

V ← ally of vocabularies

P ← ally of probability of hot news

for every title i in T

  for every vocabulary v in i

    if v exist in V

       V[v] = V[v] + 1

    else

       V[v] = 1

    end

  end

end

for every title i in T

  for every vocabulary v in i

    if v is the first four heaviest vocabularies in all vocabularies of i

      𝑃_𝑣← 0.9 * w_𝑣w_max

    end

  end

   P[i] ← 𝑃₁𝑃₂𝑃₃𝑃₄𝑃₁𝑃₂𝑃₃𝑃₄+(1−𝑃₁)(1−𝑃₂)(1−𝑃₃)(1−𝑃₄)

end

Sort P by descending

Select Top N from P

end