Table 1 Score functions. Elements of K are represented by kij, and its ith row by Ki, while positive integers i, j represent nodes (patients); VC ⊂ V represents the set of “positive” patients, i.e. patients associated with the phenotype or clinical outcome C of interest.
From: Network modeling of patients' biomolecular profiles for clinical phenotype/outcome prediction
Score Name | Formula |
|---|---|
Average | \({S}_{AV}(i,{{\boldsymbol{K}}}_{i},{V}_{C})=\frac{1}{| {V}_{C}| }{\sum }_{j\in {V}_{C}}{k}_{ij}\quad \quad (1)\) |
Nearest Neighbour | \({S}_{NN}(i,{{\boldsymbol{K}}}_{i},{V}_{C})=\mathop{max}\limits_{j\in {V}_{C}}{k}_{ij}\quad \quad (2)\) |
k-Nearest Neighbour | \(\begin{array}{ccc}{S}_{kNN}(i,{{\boldsymbol{K}}}_{i},{V}_{C}) & = & \frac{1}{|{I}_{k}(i)|}{\sum }_{j\in {I}_{k}}{k}_{ij}\\ {\rm{w}}{\rm{h}}{\rm{e}}{\rm{r}}{\rm{e}}\,{I}_{k}(i) & = & \{j|j\in {V}_{C}\wedge {\rm{r}}{\rm{a}}{\rm{n}}{\rm{k}}({k}_{ij})\le k\}\end{array}\,\,(3)\) |
Total | \({S}_{TOT}(i,{{\bf{K}}}_{i},{V}_{C})=\frac{{\sum }_{j\in {V}_{C}}{k}_{ij}}{{\sum }_{j\in {V}_{C}}{k}_{ij}+{\sum }_{j\in V\backslash {V}_{C}}{k}_{ij}}\quad \quad (4)\) |
Differential | \({S}_{Diff}(i,{{\bf{K}}}_{i},{V}_{C})={\sum }_{j\in {V}_{C}}{k}_{ij}-{\sum }_{j\in V\backslash {V}_{C}}{k}_{ij}\quad \quad (5)\) |
Differential normalized | \({S}_{Dnorm}(i,{{\bf{K}}}_{i},{V}_{C})=\frac{{\sum }_{j\in {V}_{C}}{k}_{ij}-{\sum }_{j\in V\backslash {V}_{C}}{k}_{ij}}{{\sum }_{j\in {V}_{C}}{k}_{ij}+{\sum }_{j\in V\backslash {V}_{C}}{k}_{ij}}\quad \quad (6)\) |
