House Price Prediction using Machine

Learning in Python
We all have experienced a time when we have to look up for a new house to buy. But
then the journey begins with a lot of frauds, negotiating deals, researching the local areas
and so on.

House Price Prediction using Machine Learning

So to deal with this kind of issues Today we will be preparing a MACHINE LEARNING
Based model, trained on the House Price Prediction Dataset. 
You can download the dataset from this link.
The dataset contains 13 features :
1 Id To count the records.

2 MSSubClass  Identifies the type of dwelling involved in the sale.

3 MSZoning Identifies the general zoning classification of the sale.

4 LotArea  Lot size in square feet.

5 LotConfig Configuration of the lot

6 BldgType Type of dwelling

7 OverallCond Rates the overall condition of the house

8 YearBuilt Original construction year

Remodel date (same as construction date if no remodeling or

9 YearRemodAdd additions).

1
0 Exterior1st Exterior covering on house

1
1 BsmtFinSF2 Type 2 finished square feet.
 1
2 TotalBsmtSF Total square feet of basement area

1
3 SalePrice To be predicted

Importing Libraries and Dataset

Here we are using 
 Pandas – To load the Dataframe
 Matplotlib – To visualize the data features i.e. barplot
 Seaborn – To see the correlation between features using heatmap
 Python3

import pandas as pd

import [Link] as plt

import seaborn as sns

dataset = pd.read_excel("[Link]")

# Printing first 5 records of the dataset

print([Link](5))

Output:
  

As we have imported the data. So shape method will show us the dimension of the
dataset. 

 Python3

[Link]

Output: 
(2919,13)

Data Preprocessing
Now, we categorize the features depending on their datatype (int, float, object) and then
calculate the number of them. 

 Python3

obj = ([Link] == 'object')

object_cols = list(obj[obj].index)

print("Categorical variables:",len(object_cols))
 int_ = ([Link] == 'int')

num_cols = list(int_[int_].index)

print("Integer variables:",len(num_cols))

fl = ([Link] == 'float')

fl_cols = list(fl[fl].index)

print("Float variables:",len(fl_cols))

Output: 
Categorical variables : 4
Integer variables : 6
Float variables : 3

Exploratory Data Analysis

EDA refers to the deep analysis of data so as to discover different patterns and spot
anomalies. Before making inferences from data it is essential to examine all your
variables.
So here let’s make a heatmap using seaborn library.
 Python3

[Link](figsize=(12, 6))

[Link]([Link](),

            cmap = 'BrBG',

            fmt = '.2f',

            linewidths = 2,
             annot = True)

Output:

To analyze the different categorical features. Let’s draw the barplot.

 Python3

unique_values = []

for col in object_cols:

  unique_values.append(dataset[col].unique().size)

[Link](figsize=(10,6))

[Link]('No. Unique values of Categorical Features')

[Link](rotation=90)

[Link](x=object_cols,y=unique_values)

Output:
  

The plot shows that Exterior1st has around 16 unique categories and other features have
around  6 unique categories. To findout the actual count of each category we can plot the
bargraph of each four features separately.

 Python3

[Link](figsize=(18, 36))

[Link]('Categorical Features: Distribution')

[Link](rotation=90)

index = 1

for col in object_cols:

    y = dataset[col].value_counts()
     [Link](11, 4, index)

    [Link](rotation=90)

    [Link](x=list([Link]), y=y)

    index += 1

Output:

Data Cleaning
Data Cleaning is the way to improvise the data or remove incorrect, corrupted or
irrelevant data.
As in our dataset, there are some columns that are not important and irrelevant for the
model training. So, we can drop that column before training. There are 2 approaches to
dealing with empty/null values
 We can easily delete the column/row (if the feature or record is not much important).
 Filling the empty slots with mean/mode/0/NA/etc. (depending on the dataset
requirement).
As Id Column will not be participating in any prediction. So we can Drop it.

 Python3

[Link](['Id'],

             axis=1,

             inplace=True)
 Replacing SalePrice empty values with their mean values to make the data distribution
symmetric.

 Python3

dataset['SalePrice'] = dataset['SalePrice'].fillna(

  dataset['SalePrice'].mean())

Drop records with null values (as the empty records are very less).

 Python3

new_dataset = [Link]()

Checking features which have null values in the new dataframe (if there are still any).

 Python3

new_dataset.isnull().sum()

Output:

 
 OneHotEncoder – For Label categorical features
One hot Encoding is the best way to convert categorical data into binary vectors. This
maps the values to integer values. By using OneHotEncoder, we can easily convert object
data into int. So for that, firstly we have to collect all the features which have the object
datatype. To do so, we will make a loop.
 Python3

from [Link] import OneHotEncoder

s = (new_dataset.dtypes == 'object')

object_cols = list(s[s].index)

print("Categorical variables:")

print(object_cols)

print('No. of. categorical features: ',

      len(object_cols))

Output:

Then once we have a list of all the features. We can apply OneHotEncoding to the whole
list.

 Python3

OH_encoder = OneHotEncoder(sparse=False)
 OH_cols =
[Link](OH_encoder.fit_transform(new_dataset[object_cols]))

OH_cols.index = new_dataset.index

OH_cols.columns = OH_encoder.get_feature_names()

df_final = new_dataset.drop(object_cols, axis=1)

df_final = [Link]([df_final, OH_cols], axis=1)

Splitting Dataset into Training and Testing

X and Y splitting (i.e. Y is the SalePrice column and the rest of the other columns are X)

 Python3

from [Link] import mean_absolute_error

from sklearn.model_selection import train_test_split

X = df_final.drop(['SalePrice'], axis=1)

Y = df_final['SalePrice']

# Split the training set into

# training and validation set

X_train, X_valid, Y_train, Y_valid = train_test_split(

    X, Y, train_size=0.8, test_size=0.2, random_state=0)

Model and Accuracy

 As we have to train the model to determine the continuous values, so we will be using
these regression models.
 SVM-Support Vector Machine
 Random Forest Regressor
 Linear Regressor
And To calculate loss we will be using the mean_absolute_percentage_error module. It
can easily be imported by using sklearn library. The formula for Mean Absolute Error : 

SVM – Support vector Machine

SVM can be used for both regression and classification model. It finds the hyperplane in
the n-dimensional plane. To read more about svm refer this.
 Python3

from sklearn import svm

from [Link] import SVC

from [Link] import mean_absolute_percentage_error

model_SVR = [Link]()

model_SVR.fit(X_train,Y_train)

Y_pred = model_SVR.predict(X_valid)
 print(mean_absolute_percentage_error(Y_valid, Y_pred))

Output : 
0.18705129
Random Forest Regression
Random Forest is an ensemble technique that uses multiple of decision trees and can be
used for both regression and classification tasks. To read more about random forests refer
this.
 Python3

from [Link] import RandomForestRegressor

model_RFR = RandomForestRegressor(n_estimators=10)

model_RFR.fit(X_train, Y_train)

Y_pred = model_RFR.predict(X_valid)

mean_absolute_percentage_error(Y_valid, Y_pred)

Output : 
0.1929469
Linear Regression
Linear Regression predicts the final output-dependent value based on the given
independent features. Like, here we have to predict SalePrice depending on features like
MSSubClass, YearBuilt, BldgType, Exterior1st etc. To read more about Linear
Regression refer this.
 Python3
 from sklearn.linear_model import LinearRegression

model_LR = LinearRegression()

model_LR.fit(X_train, Y_train)

Y_pred = model_LR.predict(X_valid)

print(mean_absolute_percentage_error(Y_valid, Y_pred))

Output : 
0.187416838

Conclusion 
Clearly, SVM model is giving better accuracy as the mean absolute error is the least
among all the other regressor models i.e. 0.18 approx. To get much better results
ensemble learning techniques like Bagging and Boosting can also be used.