Question 01: Predict the price of the Uber ride from a given pickup point to the agreed drop-off location. Perform following tasks:
1) Pre-process the dataset.
2) Identify outliers.
3) Check the correlation.
4) Implement linear regression and random forest regression models.
5) Evaluate the models and compare their respective scores like R2, RMSE, etc. 
Download hole Program / Project code, by clicking following link: What steps are involved in pre-processing the dataset for Uber Ride Price Prediction using Machine Learning ? Pre-processing the dataset is a crucial step for ensuring accurate model predictions. For Uber ride price prediction, the following pre-processing steps are typically performed:
1. Handling Missing Values: Check for and impute or remove any missing data.
2. Data Type Conversion: Convert columns like pickup and drop-off datetime to appropriate `datetime` format.
3. Feature Engineering: Extract useful features such as hour, day, distance (using Haversine formula), etc., from datetime and coordinate columns.
4. Removing Duplicates: Drop duplicate rows to avoid redundancy.
5. Normalization/Scaling: Apply feature scaling techniques such as StandardScaler or MinMaxScaler especially for regression algorithms.
6. Encoding Categorical Variables: Use label encoding or one-hot encoding for categorical features if present.
7. Outlier Removal: Identify and remove abnormal values in fare amounts or distances which may skew the model.
1) Pre-process the dataset.
2) Identify outliers.
3) Check the correlation.
4) Implement linear regression and random forest regression models.
5) Evaluate the models and compare their respective scores like R2, RMSE, etc.
How do Linear Regression and Random Forest Regression models compare in predicting Uber ride prices, and how are they evaluated ?
Linear Regression and Random Forest Regression are both supervised learning models but differ in complexity and performance:
Linear Regression:
Assumes a linear relationship between input features and target (ride price).
Simple to interpret and fast to train.
May underperform in presence of non-linear data or outliers.
Random Forest Regression:
Ensemble learning method using multiple decision trees.
Handles non-linear relationships well and is robust to outliers.
Typically provides higher accuracy at the cost of interpretability.
Evaluation Metrics:
R² Score (Coefficient of Determination): Indicates how well the model explains the variance in the target variable.
RMSE (Root Mean Squared Error): Measures the average magnitude of the prediction error.
Typical Result:
Random Forest generally achieves a higher R² and lower RMSE compared to Linear Regression, making it better suited for ride price prediction with complex patterns in the data.
Programming Code: Following code write in: ML_P01.py # ML Project Program 01
# import libraries
import numpy as np
import pandas as pd
# import dataset
data = pd.read_csv("uber_dataset/uber.csv")
# print first few data of uber dataset
data.head
# print information of Uber dataset
data.info()
# dtypes is nothing but the data types
# object is string data
# converting object to date & time
data["pickup_datetime"] = pd.to_datetime(data["pickup_datetime"])
data.info()
# successfully converted object to date & time by using to_datetime() method
# find missing values
data.isnull()
# find total number of missing values
data.isnull().sum()
# 0 means false & 1 means True
# if Ture means null or missing values in dataset or in row
# drop the row if it has missing values
data.dropna(inplace = True)
# After drop missing value row
data.isnull().sum()
# Now create a Machine Learning Model
# import lib
from sklearn.linear_model import LinearRegression
from sklearn.model_selection import train_test_split
from sklearn.metrics import mean_squared_error
# x is predictor variable
x = data.drop("fare_amount", axis = 1)
# y is target variable
y = data["fare_amount"]
# to apply model
x['pickup_datetime'] = pd.to_numeric(pd.to_datetime(x['pickup_datetime']))
x = x.loc[:, x.columns.str.contains('^Unnamed')]
x_train, x_test, y_train, y_test = train_test_split(x, y, test_size=0.2)
# testing data is 20%
# training data is 80%, allocated to model
# creating Linear Regression model
lrmodel = LinearRegression()
lrmodel.fit(x_train, y_train)
# model is created
# prediction
pred = lrmodel.predict(x_test)
# Calculating RMSE
lrmodelrmse = np.sqrt(mean_squared_error(pred, y_test))
print("RMSE error is: ",lrmodelrmse)
# Random Forest Regression
from sklearn.ensemble import RandomForestRegressor
# create RFR Model
rfrmodel = RandomForestRegressor(n_estimators = 100, random_state = 101)
# fit the forest
rfrmodel.fit(x_train, y_train)
rfrmodel_pred = rfrmodel.predict(x_test)
# Calculate RMSE for RFR
rfrmodel_rmse = np.sqrt(mean_squared_error(rfrmodel_pred, y_test))
print("RFR RMSE error is: ", rfrmodel_rmse)
# prediction
pred = lrmodel.predict(x_test)
print("hh",pred)
lrmodel.predict(x_test)
from sklearn import metrics
# R2 score
# R2 score Linear Regression
metrics.r2_score(y_test, pred)
# R2 score RF Model
metrics.r2_score(y_test, rfrmodel_pred)
# R2 score Linear Regression is 894% that means model not fit.
# R2 score RF Model is: 52%
# Random Forest Model best fit for this dataset, is perfect
# Thanks For Reading.
Output:
# ML Project Program 01 # import libraries import numpy as np import pandas as pd # import dataset data = pd.read_csv("uber_dataset/uber.csv") # print first few data of uber dataset data.head # print information of Uber dataset data.info() # dtypes is nothing but the data types # object is string data # converting object to date & time data["pickup_datetime"] = pd.to_datetime(data["pickup_datetime"]) data.info() # successfully converted object to date & time by using to_datetime() method # find missing values data.isnull() # find total number of missing values data.isnull().sum() # 0 means false & 1 means True # if Ture means null or missing values in dataset or in row # drop the row if it has missing values data.dropna(inplace = True) # After drop missing value row data.isnull().sum() # Now create a Machine Learning Model # import lib from sklearn.linear_model import LinearRegression from sklearn.model_selection import train_test_split from sklearn.metrics import mean_squared_error # x is predictor variable x = data.drop("fare_amount", axis = 1) # y is target variable y = data["fare_amount"] # to apply model x['pickup_datetime'] = pd.to_numeric(pd.to_datetime(x['pickup_datetime'])) x = x.loc[:, x.columns.str.contains('^Unnamed')] x_train, x_test, y_train, y_test = train_test_split(x, y, test_size=0.2) # testing data is 20% # training data is 80%, allocated to model # creating Linear Regression model lrmodel = LinearRegression() lrmodel.fit(x_train, y_train) # model is created # prediction pred = lrmodel.predict(x_test) # Calculating RMSE lrmodelrmse = np.sqrt(mean_squared_error(pred, y_test)) print("RMSE error is: ",lrmodelrmse) # Random Forest Regression from sklearn.ensemble import RandomForestRegressor # create RFR Model rfrmodel = RandomForestRegressor(n_estimators = 100, random_state = 101) # fit the forest rfrmodel.fit(x_train, y_train) rfrmodel_pred = rfrmodel.predict(x_test) # Calculate RMSE for RFR rfrmodel_rmse = np.sqrt(mean_squared_error(rfrmodel_pred, y_test)) print("RFR RMSE error is: ", rfrmodel_rmse) # prediction pred = lrmodel.predict(x_test) print("hh",pred) lrmodel.predict(x_test) from sklearn import metrics # R2 score # R2 score Linear Regression metrics.r2_score(y_test, pred) # R2 score RF Model metrics.r2_score(y_test, rfrmodel_pred) # R2 score Linear Regression is 894% that means model not fit. # R2 score RF Model is: 52% # Random Forest Model best fit for this dataset, is perfect # Thanks For Reading.
Output:
0 Comments