update model config and number display

results/auto_arima/config.json

@@ -1,5 +1,5 @@
 {
     "model": "auto_arima",
-    "model_type": "deep-learning",
+    "model_type": "statistical",
     "model_dtype": "float32"
 }

results/auto_ets/config.json

@@ -1,5 +1,5 @@
 {
     "model": "auto_ets",
-    "model_type": "deep-learning",
+    "model_type": "statistical",
     "model_dtype": "float32"
 }

results/auto_theta/config.json

@@ -1,5 +1,5 @@
 {
     "model": "auto_theta",
-    "model_type": "deep-learning",
+    "model_type": "statistical",
     "model_dtype": "float32"
 }

src/utils.py

@@ -2,6 +2,17 @@ import pandas as pd
 import os
 import re
 
+# Define the formatting function
+def format_number(num):
+    # Check if the value is numeric
+    if isinstance(num, (int, float)):
+        if abs(num) >= 10**2:
+            return f"{num:.1e}"
+        else:
+            return f"{num:.3f}"
+    # Return non-numeric values as-is
+    return num
+
 def norm_sNavie(df):
     df_normalized = df.copy()
     seasonal_naive_row = df[df['model'] == 'seasonal_naive'].iloc[0]
@@ -47,7 +58,8 @@ def pivot_existed_df(df, tab_name):
     df_pivot = df_melted.pivot_table(index='model', columns=[tab_name, 'metric'], values='value')
     df_pivot.columns = [f'{tab_name} ({metric})' for tab_name, metric in df_pivot.columns]
     df_pivot = df_pivot.reset_index()
-    df_pivot = df_pivot.round(3)
+    # df_pivot = df_pivot.round(3)
+    df_pivot = df_pivot.applymap(format_number)
     return df_pivot