kshirsagarsiddharth

import plotly.express as px 
import pandas as pd 
from dash import Dash, dcc, html, Input, Output, State
from jupyter_dash import JupyterDash
import dash_bootstrap_components as dbc 
import numpy as np
df = pd.read_csv('car_price_data.csv').drop('Unnamed: 0', axis = 1)

kshirsagarsiddharth

param_grid = {
                'regressor__alpha': [0.1,1,10,0.01,0.01,5],
                'regressor__l1_ratio': np.arange(0.40,1.00,0.10),
            }
grid_search = GridSearchCV(reg, param_grid, cv=2)

grid_search.fit(X_train, y_train)
preds = grid_search.best_estimator_.predict(X_test)
score = np.sqrt(mean_squared_error(y_test, preds))

kshirsagarsiddharth

from sklearn import set_config

set_config(display="diagram")
numeric_columns = X.select_dtypes(exclude='object').columns
numeric_transformer = Pipeline(steps = [("scalar", StandardScaler())])

unordered_columns =['transmission', 'fuel_type']
unordered_transformer = Pipeline(steps = [('onehot',  OneHotEncoder(handle_unknown='ignore', sparse=False))])

ordered_columns = ['year', 'model']

kshirsagarsiddharth

elastic_net_parag_grid =  {
                'alpha'     : [0.1,1,10,0.01,0.01,5],
                'l1_ratio'  :  np.arange(0.40,1.00,0.10),
            }
elastic_net_regressor = ElasticNet()
elastic_net_grid_search = GridSearchCV(
    elastic_net_regressor,
    param_grid=elastic_net_parag_grid,
    cv = 5,
    n_jobs=-1,

kshirsagarsiddharth

### Function to test various models
def score_dataset(X_train,X_valid,y_train,y_valid, input_model):
    model = input_model
    model.fit(X_train,y_train)
    preds = model.predict(X_valid)
    return np.sqrt(mean_squared_error(y_valid, preds))


input_model = LinearRegression()
score_dataset(encoded_X_train,encoded_X_valid,y_train,y_valid, input_model)

kshirsagarsiddharth

def transform_columns(transformer,columns_to_transform,X_train = X_train,X_valid = X_valid,):
    cols_train =  pd.DataFrame(transformer.fit_transform(X_train[columns_to_transform]))
    cols_valid = pd.DataFrame(transformer.transform(X_valid[columns_to_transform]))
    cols_train.index = X_train.index
    cols_valid.index = X_valid.index
    return cols_train, cols_valid

columns_to_transform_list = [ordered_columns, unordered_columns, numeric_columns]
ordinal_encoder = OrdinalEncoder()
onehot_encoder = OneHotEncoder(handle_unknown='ignore', sparse=False)

kshirsagarsiddharth

# Add one-hot encoded columns and numerical features
encoded_X_train = pd.concat([oh_cols_train, sc_cols_train,ord_cols_train], axis = 1)
encoded_X_valid = pd.concat([oh_cols_valid, sc_cols_valid,ord_cols_valid], axis = 1)
encoded_X_train

kshirsagarsiddharth

# Remove categorical columns (will replace with one-hot encoding)
sc = StandardScaler()
sc_cols_train = pd.DataFrame(sc.fit_transform(X_train[numeric_columns]))
sc_cols_valid = pd.DataFrame(sc.transform(X_valid[numeric_columns]))
# One-hot encoding removed index; put it back
sc_cols_train.index = X_train.index
sc_cols_valid.index = X_valid.index

kshirsagarsiddharth

ord_encoder = OrdinalEncoder()
ord_cols_train = pd.DataFrame(ord_encoder.fit_transform(X_train[ordered_columns]))
ord_cols_valid = pd.DataFrame(ord_encoder.transform(X_valid[ordered_columns]))
ord_cols_train.index = X_train.index
ord_cols_valid.index = X_valid.index

kshirsagarsiddharth

oh_encoder = OneHotEncoder(handle_unknown='ignore', sparse=False)
oh_cols_train = pd.DataFrame(oh_encoder.fit_transform(X_train[unordered_columns]))
oh_cols_valid = pd.DataFrame(oh_encoder.transform(X_valid[unordered_columns]))
# One-hot encoding removed index; put it back
oh_cols_train.index = X_train.index
oh_cols_valid.index = X_valid.index