From fbb5e9f094d5346d22918ddf291f92bcfa3b6a9f Mon Sep 17 00:00:00 2001
From: Shreyash Somvanshi <91454227+ShreyashSomvanshi@users.noreply.github.com>
Date: Thu, 11 May 2023 17:42:26 +0530
Subject: [PATCH] Create 550. Choose a better regression model.py

Correct but not getting accepted even after passing test cases
---
 .../550. Choose a better regression model.py  | 44 +++++++++++++++++++
 1 file changed, 44 insertions(+)
 create mode 100644 Machine Learning/550. Choose a better regression model.py

diff --git a/Machine Learning/550. Choose a better regression model.py b/Machine Learning/550. Choose a better regression model.py
new file mode 100644
index 0000000..5b64747
--- /dev/null
+++ b/Machine Learning/550. Choose a better regression model.py	
@@ -0,0 +1,44 @@
+'''
+A petroleum refining company wants to predict the price of gasoline based on different correlated 
+factors like the cost of crude oil, refining cost, marketing cost, fuel tax, etc. You are given the
+following training and testing data as inputs containing different feature values. Fit a linear regression
+model and a bayesian regression model using the given training data. Once fitted, compare the 
+mean squared error value between the models and return the best-fitted regression model based on 
+the minimum mean squared error value.
+
+Example 1:
+Input:
+X_train=[[10.2,9],[18.2,3.3],[14,13],[19,14],[12,23],[16,17]]
+X_test=[[10,16],[22,15],[14,23],[22,26]]
+y_train=[22.2,44.3,22.4,22.6,34,21.3]
+y_test=[22,23,21,20]
+Output: LinearRegression
+
+Example 2:
+Input:
+X_train=[[59,19],[57.7,3],[56,5],[60,4],[52,3],[64,7]]
+X_test=[[60,6],[62,5],[54,3],[62,6]]
+y_train=[52,44,42.4,42.6,44,41.3]
+y_test=[62,63,61,60]
+Output: BayesianRidge
+'''
+
+# Solution:
+from sklearn.linear_model import LinearRegression, BayesianRidge
+from sklearn.metrics import mean_squared_error
+def MeanAbsoluteError(X_train,X_test,y_train,y_test):
+  lr = LinearRegression()
+  lr.fit(X_train, y_train)
+  lr_pred = lr.predict(X_test)
+  lr_err = mean_squared_error(y_test, lr_pred)
+  br = BayesianRidge()
+  br.fit(X_train, y_train)
+  br_pred = br.predict(X_test)
+  br_err = mean_squared_error(y_test, br_pred)
+  if(br_err < lr_err):
+    print('BaysianRidge')
+  else:
+    print('LinearRegression')
+  
+MeanAbsoluteError(X_train,X_test,y_train,y_test)  
+