Background
Data set
Proposed framework
Data preprocessing
Similarity measure for time series
Euclidean distance
Dynamic time warping
Triangle distance metric
Hierarchical cluster analysis
AGNES algorithm | Cophenetic correlation coefficient (CPCC) | ||
---|---|---|---|
DTW | TDM | Euclidean | |
Single | 0.7378 | 0.7101 | 0.6650 |
Complete | 0.5690 | 0.5124 | 0.5266 |
Average |
0.7949
|
0.7692
|
0.7411
|
Ward | 0.6003 | 0.6306 | 0.6199 |
Weighted | 0.7186 | 0.6916 | 0.6886 |
Median | 0.6602 | 0.6284 | 0.5627 |
Centroid | 0.7630 | 0.7243 | 0.6872 |
Time series merging
Trend analysis
Results and discussion
Cluster analysis
Gujrat_cluster | |
---|---|
Cluster1 | Ahmedabad, Surat, Rajkot, Bansakantha, Junagadh, Panch Mahals, Vadodara, Sabar Kantha |
Cluster2 | Tapi, Narmada, Porbandar, Kachch, Surendranagar, Gandhi Nagar, Dahod, Amreli, Jamnagar, Anand, Bharuch, Bhavnagar, Kheda, Mahasena, Patan, Navsari, Valsad |
Cluster3 | The Dangs |
Uttarakhand_cluster | |
---|---|
Cluster1 | Dehradun, Udhamsinghnagar, Naintal, |
Cluster2 | Almora, Pauri, Pithoragarh |
Cluster3 | Tehri, Chamoli, Uttarkashi, Rudraprayag, Bageshwar, Champawat |
Cluster4 | Haridwar |
Trend analysis
Gujrat clusters | Uttarakhand cluster | ||||||
---|---|---|---|---|---|---|---|
T.S. Id | C1 | C2 | C3 | C1 | C2 | C3 | C4 |
1 | N | P | P | P | P | P | N |
2 | P | P | – | P | N | N | – |
3 | N | P | – | P | P | N | – |
4 | N | P | – | – | – | P | – |
5 | P | P | – | – | – | P | – |
6 | N | N | – | – | – | P | – |
7 | P | P | – | – | – | – | – |
8 | N | P | – | – | – | – | – |
9 | – | P | – | – | – | – | – |
10 | – | P | – | – | – | – | – |
11 | – | P | – | – | – | – | – |
12 | – | P | – | – | – | – | – |
13 | – | P | – | – | – | – | – |
14 | – | N | – | – | – | – | – |
15 | – | N | – | – | – | – | – |
16 | – | N | – | – | – | – | – |
17 | – | P | – | – | – | – | – |
Total P | 3 | 13 | 1 | 3 | 2 | 4 | 0 |
Total N | 5 | 4 | 0 | 0 | 1 | 2 | 1 |
PTMA | N | P | P | P | P | P | N |
ATMA | P | N | P | P | P | N | N |