Table 4.
Details of transfer functions used in multilayer perceptron.
| Acronym | Long Term | Equations |
|---|---|---|
| Elliotsig | Elliot sigmoid | f(x) = elliotsig(x) = (x)/1+|x| |
| Hardlim | Positive hard limit |
f(x) = hardlim(x) = 1, if x ≥ 0;
=0, otherwise. |
| Hardlims | Symmetric hard limit |
f(x) = hardlim(x) = 1, if x ≥ 0;
=−1, otherwise. |
| Logsig | Logarithmic sigmoid | f(x) = logsig(x) = 1/(1 + exp(−x)) |
| Netinv | Inverse | f(x) = netinv (x) = 1/x |
| Poslin | Positive linear |
f(x) = poslin(x) = x, if x
≥
0;
=0, if x ≤ 0. |
| Purelin | Linear | f(x) = purelin(x) = x |
| Radbas | Radial basis | f(x) = radbas(x) = exp (−x2) |
| Radbasn | Radial basis normalized | f(x) = radbasn(x) = exp (−x2)/sum(exp(−x2)) |
| Satlin | Positive saturating linear |
f(x) = satlin(x) = 0, if x ≤ 0;
=x, if 0 ≤ x ≤ 1; =1, if 1 ≤ x. |
| Satlins | Symmetric saturating linear |
f(x) = satlins(x) = −1, if x ≤ −1;
=x, if −1 ≤ x ≤ 1; =1, if 1 ≤ x. |
| Softmax | Soft max | f(x) = softmax(x) = exp(x)/sum(exp(x)) |
| Tansig | Symmetric sigmoid | f(x) = tansig(x) = 2/(1+exp (−2 ∗ x)) − 1 |
| Tribas | Triangular basis |
f(x) = tribas(x) = 1 − abs(x), if −1 ≤ x ≤ 1;
=0, otherwise. |
Where x is the weighted sum of w, i, b, and y of Equation (2).