Table 4 \(f\left(x\right)\) implementation steps.
From: An efficient bearing fault detection strategy based on a hybrid machine learning technique
\(f\left(x\right)\) | ||
|---|---|---|
\(Define the initial hyperparameters set {C}_{0}\) \(1. Define the search space according to Table 2.\) \(2. Initialize: {v}^{0}=ac{c}_{val}^{0}-\lambda * {t}_{val}^{0}.\) \(D=\left\{\left({c}^{0},{v}^{0}\right)\right\},{v}_{best}\) \({v}^{0} {c}_{best}={c}^{0}\) \(3. for i \leftarrow 1 to {n}_{Iter} do\) \(calculat l\left(c\right) { D}_{good}=\left\{\left(c , v\right) \mid v {>v}^{*}\right\}\) \({calculat l\left(c\right) D}_{bad}=\left\{\left(c , v\right) \mid v {\le v}^{*}\right\}\) \(update: {v}^{*}=\gamma \times {v}_{best}\) \(choose the new hyperparameters set: {c}^{i+1}= {c}^{0} =argmax\frac{l\left(c\right)}{g(c)}\) \(Calculate the new evaluation value: {v}^{i+1}=ac{c}_{val}^{i+1}-\lambda \times {t}_{val}^{t+1}\) \(4. end\) |
