kozy9 commited on
Commit
08faf09
·
verified ·
1 Parent(s): 5c0af90

Upload SARIMAX(2, 1, 1)x(2, 0, 2, 12) — test RMSE 5.1561

Browse files
README.md CHANGED
@@ -17,7 +17,7 @@ using historical water level data and meteorological variables.
17
 
18
  | Parameter | Value |
19
  |---|---|
20
- | Architecture | SARIMAX(1, 1, 2)x(1, 0, 2, 12) |
21
  | Seasonal period | 12 months |
22
  | Target | Water level (m) |
23
  | Exogenous variables | Temperature (°C), Precipitation (mm), Wind Speed (km/h) |
@@ -27,23 +27,20 @@ using historical water level data and meteorological variables.
27
 
28
  ## Hyperparameter Tuning
29
 
30
- Grid search over 243 combinations,
31
  ranked by out-of-sample validation RMSE.
32
 
33
- | Criterion | Value |
34
- |---|---|
35
- | Best validation RMSE | 3.9392 |
36
- | Best AIC | 4624.6238 |
37
 
38
  ## Test Set Performance
39
 
40
  | Metric | Value |
41
  |---|---|
42
- | RMSE | 5.2523 |
43
- | MAE | 4.2746 |
44
- | MAPE (%) | 6.7072 |
45
- | R² | -0.4359 |
46
- | NSE | -0.4359 |
47
 
48
  > This model is a statistical baseline for benchmarking against
49
  > deep learning approaches (LSTM, TCN).
 
17
 
18
  | Parameter | Value |
19
  |---|---|
20
+ | Architecture | SARIMAX(2, 1, 1)x(2, 0, 2, 12) |
21
  | Seasonal period | 12 months |
22
  | Target | Water level (m) |
23
  | Exogenous variables | Temperature (°C), Precipitation (mm), Wind Speed (km/h) |
 
27
 
28
  ## Hyperparameter Tuning
29
 
30
+ Bayesian search : 50 trials | criterion: validation RMSE
31
  ranked by out-of-sample validation RMSE.
32
 
33
+
 
 
 
34
 
35
  ## Test Set Performance
36
 
37
  | Metric | Value |
38
  |---|---|
39
+ | RMSE | 5.1561 |
40
+ | MAE | 4.1987 |
41
+ | MAPE (%) | 6.5872 |
42
+ | R² | -0.3838 |
43
+ | NSE | -0.3838 |
44
 
45
  > This model is a statistical baseline for benchmarking against
46
  > deep learning approaches (LSTM, TCN).
fig4_forecast.png CHANGED

Git LFS Details

  • SHA256: a46987237920920189e83bc5e18d5b8cc7f449b35f48c67e50d48e17fa3fa1a1
  • Pointer size: 131 Bytes
  • Size of remote file: 135 kB

Git LFS Details

  • SHA256: d27e129cd093517613abfb09a10f40fda9cc35a36934a1306a8965744f92edc7
  • Pointer size: 131 Bytes
  • Size of remote file: 134 kB
fig5_error_analysis.png CHANGED
fig6_diagnostics.png CHANGED

Git LFS Details

  • SHA256: a1f9021a2c01634f350358c4dd73a775073e7ab7cb669bd46d2e80a48d571b26
  • Pointer size: 131 Bytes
  • Size of remote file: 166 kB

Git LFS Details

  • SHA256: c0f6a16267ed3fec084288ea1374398feca0c943ed4c4c7e43b6a79cc0f368e0
  • Pointer size: 131 Bytes
  • Size of remote file: 168 kB
model_config.json CHANGED
@@ -1,12 +1,12 @@
1
  {
2
  "model_type": "SARIMAX",
3
  "order": [
 
4
  1,
5
- 1,
6
- 2
7
  ],
8
  "seasonal_order": [
9
- 1,
10
  0,
11
  2,
12
  12
@@ -31,17 +31,18 @@
31
  "n_obs": 96
32
  },
33
  "test_metrics": {
34
- "RMSE": 5.2523,
35
- "MAE": 4.2746,
36
- "MAPE_pct": 6.7072,
37
- "R2": -0.4359,
38
- "NSE": -0.4359
39
  },
40
- "grid_search": {
41
- "combinations_tested": 243,
 
42
  "criterion": "val_RMSE",
43
- "best_val_RMSE": 3.9392,
44
- "best_AIC": 4624.6238
45
  },
46
  "notes": "Basic SARIMAX \u2014 raw exog only, no lag features. Oracle exog assumption at forecast time."
47
  }
 
1
  {
2
  "model_type": "SARIMAX",
3
  "order": [
4
+ 2,
5
  1,
6
+ 1
 
7
  ],
8
  "seasonal_order": [
9
+ 2,
10
  0,
11
  2,
12
  12
 
31
  "n_obs": 96
32
  },
33
  "test_metrics": {
34
+ "RMSE": 5.1561,
35
+ "MAE": 4.1987,
36
+ "MAPE_pct": 6.5872,
37
+ "R2": -0.3838,
38
+ "NSE": -0.3838
39
  },
40
+ "bayesian_optimization": {
41
+ "n_trials": 50,
42
+ "sampler": "Optuna TPE",
43
  "criterion": "val_RMSE",
44
+ "best_val_RMSE": 3.9717,
45
+ "best_AIC": 4623.9462
46
  },
47
  "notes": "Basic SARIMAX \u2014 raw exog only, no lag features. Oracle exog assumption at forecast time."
48
  }
sarimax_model.pkl CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:bfe578b4f1ce192b69f2475e1adbf9cdd32ec2e49340865d4c7fa972b58b5050
3
- size 130875715
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c892e377360030ae904b9b068dc358a6f4a4ed9e690069efb0e747b62d91a6d9
3
+ size 121925408