final attempt

run.py

@@ -19,9 +19,6 @@ dev_expected = pd.read_csv("dev-0/expected.tsv", header=None, sep="\t")
 dev_0 = pd.read_csv("dev-0/in.tsv", header=None, sep="\t")
 test_A = pd.read_csv("test-A/in.tsv", header=None, sep="\t")
 
-poly = PolynomialFeatures(2, interaction_only=True)
-
-
 def preprocess_data(df_to_process, main_df=None):
     final_df = pd.get_dummies(df_to_process, columns=[0, 3, 4])
     final_df.drop(columns=[1, 2], inplace=True)

run2.py

@@ -5,6 +5,20 @@ from sklearn.linear_model import LinearRegression
 from tensorflow.keras.models import Sequential
 from tensorflow.keras.layers import Dense
 
+import numpy as np
+import pandas as pd
+import xgboost as xg
+from sklearn.compose import TransformedTargetRegressor
+from sklearn.model_selection import GridSearchCV
+from sklearn.preprocessing import (
+    QuantileTransformer,
+    StandardScaler,
+    PolynomialFeatures,
+)
+
+import tensorflow.keras
+from tensorflow.keras.models import Sequential
+from tensorflow.keras.layers import Dense
 # Import the required library
 from geopy.geocoders import Nominatim
 
@@ -20,30 +34,32 @@ df_test = pd.read_csv("dev-0/in.tsv", names=in_columns, sep="\t")
 df = pd.concat([df, df_test])
 # df = df.drop(["nazwa_stacji"], axis=1)
 x = pd.get_dummies(df, columns=["id_stacji", "rok", "miesiąc"])
-geo_lat = {
-    "BIEBRZA-PIEŃCZYKÓWEK" : 53.65
-}
-geo_long = {
-    "BIEBRZA-PIEŃCZYKÓWEK": 22.58
-}
-for xd in x["nazwa_stacji"].unique():
-    location = geolocator.geocode(xd)
-    if xd == "BIEBRZA-PIEŃCZYKÓWEK":
-        pass
-    else:
-        print(xd)
-        geo_lat[xd] = location.latitude
-        geo_long[xd] = location.longitude
-
-
-x["latitude"] = x["nazwa_stacji"].map(geo_lat)
-x["longitude"] = x["nazwa_stacji"].map(geo_long)
+# geo_lat = {
+#     "BIEBRZA-PIEŃCZYKÓWEK" : 53.65
+# }
+# geo_long = {
+#     "BIEBRZA-PIEŃCZYKÓWEK": 22.58
+# }
+# for xd in x["nazwa_stacji"].unique():
+#     location = geolocator.geocode(xd)
+#     if xd == "BIEBRZA-PIEŃCZYKÓWEK":
+#         pass
+#     else:
+#         print(xd)
+#         geo_lat[xd] = location.latitude
+#         geo_long[xd] = location.longitude
+#
+#
+# x["latitude"] = x["nazwa_stacji"].map(geo_lat)
+# x["longitude"] = x["nazwa_stacji"].map(geo_long)
 x = x.drop(["nazwa_stacji", "typ_zbioru"], axis=1)
-
-print(x)
-print(geo_lat)
-print(geo_long)
+poly = PolynomialFeatures(2, interaction_only=True)
+#
+# print(x)
+# print(geo_lat)
+# print(geo_long)
 x = x.iloc[:-600]
+x = poly.fit_transform(x)
 y = pd.read_csv("train/expected.tsv", sep="\t", names=["rainfall"])
 
 
@@ -51,59 +67,90 @@ from sklearn.preprocessing import PolynomialFeatures
 # xxx
 # poly = PolynomialFeatures(2, interaction_only=True)
 # df = poly.fit_transform(x)
+param_grid = {
+    "n_estimators": [100, 80, 60, 55, 51, 45],
+    "max_depth": [7, 8],
+    "reg_lambda": [0.26, 0.25, 0.2],
+}
 
-model = Sequential(
-    [
-        Dense(512, activation="relu", input_dim=75),
-        tensorflow.keras.layers.BatchNormalization(),
-        Dense(512 // 2, activation="relu"),
-        tensorflow.keras.layers.BatchNormalization(),
-        Dense(512 // 4, activation="relu"),
-        tensorflow.keras.layers.BatchNormalization(),
-        Dense(512 // 8, activation="relu"),
-        tensorflow.keras.layers.BatchNormalization(),
-        Dense(32, activation="relu"),
-        tensorflow.keras.layers.BatchNormalization(),
-        Dense(1),
-    ]
+grid = GridSearchCV(
+    xg.XGBRFRegressor(), param_grid, refit=True, verbose=3, n_jobs=-1
+)  #
+regr_trans = TransformedTargetRegressor(
+    regressor=grid, transformer=QuantileTransformer(output_distribution="normal")
 )
 
-model.compile(
-    loss="mean_squared_error", optimizer="adam", metrics=["mean_squared_error"]
+# fitting the model for grid search
+grid_result = regr_trans.fit(x, y)
+best_params = grid_result.regressor_.best_params_
+
+# using best params to create and fit model
+best_model = xg.XGBRFRegressor(
+    max_depth=best_params["max_depth"],
+    n_estimators=best_params["n_estimators"],
+    reg_lambda=best_params["reg_lambda"],
 )
-model.fit(x, y, epochs=100)
+regr_trans = TransformedTargetRegressor(
+    regressor=best_model, transformer=QuantileTransformer(output_distribution="normal")
+)
+
+regr_trans.fit(x, y)
+
+# model = Sequential(
+#     [
+#         Dense(512, activation="relu", input_dim=75),
+#         tensorflow.keras.layers.BatchNormalization(),
+#         Dense(512 // 2, activation="relu"),
+#         tensorflow.keras.layers.BatchNormalization(),
+#         Dense(512 // 4, activation="relu"),
+#         tensorflow.keras.layers.BatchNormalization(),
+#         Dense(512 // 8, activation="relu"),
+#         tensorflow.keras.layers.BatchNormalization(),
+#         Dense(32, activation="relu"),
+#         tensorflow.keras.layers.BatchNormalization(),
+#         Dense(1),
+#     ]
+# )
+#
+# model.compile(
+#     loss="mean_squared_error", optimizer="adam", metrics=["mean_squared_error"]
+# )
+# model.fit(x, y, epochs=100)
 
 x_test = pd.read_csv("test-A/in.tsv", sep="\t", names=in_columns)
 df_train = pd.read_csv("train/in.tsv", names=in_columns, sep="\t")
 
-geo_lat = {
-    "BIEBRZA-PIEŃCZYKÓWEK" : 53.65
-}
-geo_long = {
-    "BIEBRZA-PIEŃCZYKÓWEK": 22.58
-}
+# geo_lat = {
+#     "BIEBRZA-PIEŃCZYKÓWEK" : 53.65
+# }
+# geo_long = {
+#     "BIEBRZA-PIEŃCZYKÓWEK": 22.58
+# }
 x_test = pd.concat([x_test, df_train])
 
-for xd in x_test["nazwa_stacji"].unique():
-    location = geolocator.geocode(xd)
-    if xd == "BIEBRZA-PIEŃCZYKÓWEK":
-        pass
-    else:
-        print(xd)
-        geo_lat[xd] = location.latitude
-        geo_long[xd] = location.longitude
-
-
-x_test["latitude"] = x_test["nazwa_stacji"].map(geo_lat)
-x_test["longitude"] = x_test["nazwa_stacji"].map(geo_long)
+# for xd in x_test["nazwa_stacji"].unique():
+#     location = geolocator.geocode(xd)
+#     if xd == "BIEBRZA-PIEŃCZYKÓWEK":
+#         pass
+#     else:
+#         print(xd)
+#         geo_lat[xd] = location.latitude
+#         geo_long[xd] = location.longitude
+#
+#
+# x_test["latitude"] = x_test["nazwa_stacji"].map(geo_lat)
+# x_test["longitude"] = x_test["nazwa_stacji"].map(geo_long)
 
 x_test = x_test.drop(["nazwa_stacji", "typ_zbioru"], axis=1)
 x_test = pd.get_dummies(x_test, columns=["id_stacji", "rok", "miesiąc"])
-
+poly = PolynomialFeatures(2, interaction_only=True)
 x_test = x_test.iloc[:-8760]
+x_test = poly.fit_transform(x_test)
 # poly = PolynomialFeatures(2, interaction_only=True)
 # x_test2 = poly.fit_transform(x_test)
-pred = model.predict(x_test)
+# pred = model.predict(x_test)
+test_A_predicted = regr_trans.predict(x_test)
 
-out = pd.DataFrame(pred)
+
+out = pd.DataFrame(test_A_predicted)
 out.to_csv("test-A/out.tsv", sep="\t", header=False, index=False)

run3.py

@@ -7,8 +7,10 @@ import pandas as pd
 import xgboost as xg
 
 import tensorflow.keras
+from keras.layers import Dropout
 from tensorflow.keras.models import Sequential
 from tensorflow.keras.layers import Dense
+from tensorflow.python.keras.wrappers.scikit_learn import KerasRegressor
 
 in_columns = ["id_stacji", "nazwa_stacji", "typ_zbioru", "rok", "miesiąc"]
 df = pd.read_csv("train/in.tsv", names=in_columns, sep="\t")
@@ -20,28 +22,44 @@ x = x.drop(["nazwa_stacji", "typ_zbioru"], axis=1)
 x = x.iloc[:-600]
 y = pd.read_csv("train/expected.tsv", sep="\t", names=["rainfall"])
 
+
 model = Sequential(
     [
-        Dense(1024, activation="relu", input_dim=73),
-        Dense(512, activation="relu"),
+        Dense(2048, activation="relu", input_dim=73),
+        Dense(1024, activation="relu"),
+        Dense(224, activation="relu"),
         # tensorflow.keras.layers.BatchNormalization(),
-        Dense(512 // 2, activation="relu"),
+        Dense(320, activation="relu"),
         # tensorflow.keras.layers.BatchNormalization(),
-        Dense(512 // 4, activation="relu"),
+        Dense(384, activation="relu"),
         # tensorflow.keras.layers.BatchNormalization(),
-        Dense(512 // 8, activation="relu"),
+        Dense(416, activation="relu"),
         # tensorflow.keras.layers.BatchNormalization(),
-        Dense(32, activation="relu"),
+        Dense(448, activation="relu"),
+        Dense(448, activation="relu"),
+        Dense(256, activation="relu"),
         # tensorflow.keras.layers.BatchNormalization(),
         Dense(1, activation="linear"),
     ]
 )
 
+# input = tensorflow.keras.layers.Input(shape=x.shape[1:])
+# hidden1 = tensorflow.keras.layers.Dense(1024, activation='relu')(input)
+# hidden2 = tensorflow.keras.layers.Dense(512, activation='relu')(hidden1)
+# hidden3 = tensorflow.keras.layers.Dense(256, activation='relu')(hidden2)
+# hidden4 = tensorflow.keras.layers.Dense(128, activation='relu')(hidden3)
+# concat = tensorflow.keras.layers.Concatenate()([input, hidden4])
+# output = tensorflow.keras.layers.Dense(1, activation="linear")(concat)
+# model = tensorflow.keras.models.Model(inputs=[input], outputs=[output])
+
 model.compile(
     loss="mean_squared_error", optimizer="adam", metrics=["mean_squared_error"]
 )
-model.fit(x, y, epochs=100)
 
+# estimator = KerasRegressor(build_fn=model, epochs=100, batch_size=10, verbose=0)
+# estimator.fit(x, y)
+model.fit(x, y, epochs=100)
+# exit()
 x_test = pd.read_csv("test-A/in.tsv", sep="\t", names=in_columns)
 df_train = pd.read_csv("train/in.tsv", names=in_columns, sep="\t")
 
@@ -50,7 +68,7 @@ x_test = pd.concat([x_test, df_train])
 x_test = x_test.drop(["nazwa_stacji", "typ_zbioru"], axis=1)
 x_test = pd.get_dummies(x_test, columns=["id_stacji", "rok", "miesiąc"])
 x_test = x_test.iloc[:-8760]
-
 pred = model.predict(x_test)
+# pred = estimator.predict(x_test)
 out = pd.DataFrame(pred)
-out.to_csv("test-A/out.tsv", sep="\t", header=False, index=False)
+out.to_csv("test-A/out2.tsv", sep="\t", header=False, index=False)