Algorithm 2: NADCA-L algorithm.

Given a sample i, U, a set of NM measures [$M_{i-NM-1}$…$M_i$], a set of K predictions [${\Delta }_{i-K}$…${\Delta }_{i-1}$] and prediction errors [$C_{i-K}$…$C_{i-1}$] for the set of measures [$M_{i-K}$…$M_{i-1}$] and $M_{i+1}$:  1. Calculate the set of differences PD using the NM measures.  2. Calculate ${\Delta }_i$ using PM and PD.  3. Calculate $C_i^{*}$ and $C_i^{**}$.  4. Calculate $P_{i+1}^{Sensor}$ and $P_{i+1}^{Sensor}{}^{**}$ using (4).  5. Calculate $dis1=|M_{i+1}^{Sensor}-P_{i+1}^{Sensor}|$ and $dis2=\left|M_{i+1}^{Sensor}-P_{i+1}^{Sensor}\left|+\right|C_i^{*}-C_i^{**}\right|$  6. If $dis1\le U$ and $dis2\le U$ then “No anomaly” at i + 1. Save (${\Delta }_i$, $C_i$) for the next iteration. Updating K ←K+1 allows the same IM to be used for the next iteration.  7. If $dis1>U$ then “fast-changing anomaly” at i + 1. Correct the anomaly at i + 1 changing $M_{i+1}$ to $P_{i+1}$. Save (${\Delta }_i$, $C_i$) for the next iteration. Updating K ←K+1 allows the same IM to be used for the next iteration.  8. If $dis2>U$ and $dis1<U$ then “slow-moving anomaly” at i + 1. Correct the anomaly at i + 1 changing $M_{i+1}$ to $P_{i+1}$. Save (${\Delta }_i$, $C_i$) for the next iteration. Updating K ←K+1 allows the same IM to be used for the next iteration.