Table 2 The algorithm for the label spreading semi-supervised method (Zhou et al. 2003)

1. Form the affinity matrix ω defined by ω_ij=exp(−|x_i−x_j|²/2σ²) if i≠j and ω_ii=0.

2. Construct the matrix S=D^−1/2ωD^−1/2 in which D is a diagonal matrix with its (i,i)-element equal to the sum of the i^th row of ω.

3. Iterate F(r+1)=βSF(r)+(1−β)Y until convergence where β is a parameter in (0, 1).

4. Let F^∗ denote the limit of the sequence {F(r)}. Label each point x_i as a label \(y_{i} =argmax_{j \leq c} \tilde {F}^{*}_{{ij}}\).