Fig. 1
From: AliClu - Temporal sequence alignment for clustering longitudinal clinical data
The proposed AliClu approach. First, raw data is pre-processed to obtain PE sequences. Then, pairwise sequence alignment is performed and a similarity matrix S is obtained. Next, S is converted into a distance matrix D. Agglomerative clustering is then performed with this distance matrix D. Validation of the clustering results is accomplished via a bootstrapping approach. In the end, retrieved clusters are analysed by the clinicians