A

Report on
Automation of Price Prediction Using Machine
Learning in a Large Furniture Company

Rudrakash Chaudhari B-112

Hrishikesh Saindane B-113

Aniket Satiate B-118

Sarthak Pandit B-92

Prathamesh Thakare B-132

Under the Guidance of

Ms. R.N. Malage

Assistant Professor, AI & DS Department

[TY BTech. Artificial Intelligence & Data Science]

AY (2024-2025)

At

DEPARTMENT OF ARTIFICIAL INTELLIGENCE & DATA SCIENCE

AISSMS IOIT Pune

Affiliated to

SPPU
 CERTIFICATE
This is to certify that the report entitled “Automation of Price

Prediction Using Machine Learning in a Large Furniture Company”

submitted by students Rudrakash Chaudhari, Hrishikesh Saindane, Aniket Satiate,

Sarthak Pandit, Prathamesh Thakre are a record of work carried out by them, in

partial fulfillment of requirement for Third Year of Engineering AI-DS at AISSMS

IOIT Pune 411001. This work is done during the academic year 2024-2025, under

our guidance.

Date:
Place:
Pune

Guide
[Ms.R.N.Malage]
 Flow of Report

Title page/Certificate/Acknowledgement
Index
Abstract
Hardware /Software Requirement
Methodology
Modules if any
Working
Features
Connectivity code (front end with back end)
Test Cases
Advantages/Limitation
Future Scope
Conclusion
References
Abstract
Price prediction has become a critical aspect of decision-making in the
furniture industry, especially in large companies with diverse product lines.
This project focuses on automating price prediction using machine learning
models to ensure accuracy, efficiency, and scalability. By leveraging historical
sales data, material costs, and market demand, machine learning algorithms can
predict prices with greater accuracy, thus helping companies optimize their
pricing strategy. This report outlines the complete process, from data
preprocessing to model evaluation, along with the hardware and software
requirements, working methodology, advantages, and future scope.

Hardware/Software Requirement
1. Hardware Requirements

To effectively run the machine learning models for price prediction, the
following hardware components are required:
• Computer/Server Specifications:
◦ Processor: Intel Core i5 or higher
◦ RAM: Minimum 8 GB (recommended 16 GB or higher for large
datasets)
◦ Storage: 500 GB HDD/SSD
◦ GPU: Optional, required for deep learning models
◦ Network Connectivity: High-speed internet for cloud operations and
database access
• Data Storage System:
◦ Cloud storage (e.g., AWS S3, Google Cloud Storage) or a local
storage server to handle large datasets.
2. Software Requirements

• Operating System:
◦ Windows, Linux, or macOS
• Programming Languages:
 ◦ Python: Preferred for machine learning tasks due to its robust
ecosystem of libraries.

•
• Libraries and Frameworks:
◦ Pandas: For data manipulation and preprocessing.
◦ Scikit-learn: For building machine learning models such as linear
regression, decision trees, and random forest.
◦ TensorFlow/PyTorch: For deep learning models (optional).
◦ Matplotlib/Seaborn: For data visualization and reporting.
◦ SQL: For database management, if data is stored in relational
databases.
• Development Environment:
◦ Jupyter Notebook or PyCharm for Python development.
• Version Control:
◦ Git for code version control.
◦ GitHub/GitLab for code repository management.
• Database:
◦ MySQL, PostgreSQL, or SQLite to store and retrieve historical
sales data and other information.

Methodology
The methodology for the automation of price prediction in a large furniture
company involves the following steps:
Data Collection

• Historical Sales Data: Collect sales records of various furniture items,

including product category, dimensions, materials used, production cost,
and sales price.
• Market Demand: Integrate external data sources (such as market trends,
customer behavior, and regional demand) to factor into pricing.
• Material Cost: Factor in fluctuating material costs based on suppliers’
prices and current market rates.
Data Preprocessing

• Data Cleaning: Handle missing values, remove duplicates, and address

inconsistencies.
• Feature Selection: Select relevant features for the model, such as product
material, size, production cost, market demand, and competitors' prices.
• Data Normalization: Normalize numerical values to improve the
performance of the machine learning model.
• Train-Test Split: Split the dataset into training and testing sets (typically
80% for training and 20% for testing).
Model Selection

• Linear Regression: Suitable for a simple price prediction model based on

linear relationships.
• Random Forest/Decision Trees: Used for more complex models with
non-linear dependencies between features.
• Deep Learning (optional): For more intricate patterns, neural networks
can be applied, but they may require more computational resources.
Model Training

• The selected model is trained on the preprocessed data. During training,

the model learns the relationship between input features (e.g., material
cost, demand forecast) and the target variable (price).
Model Evaluation

• After training, the model is tested on the reserved test dataset. Evaluation
metrics like Mean Absolute Error (MAE), Mean Squared Error
(MSE), and R² Score are used to assess performance.
Model Tuning

• Hyperparameter tuning is performed to optimize the model's

performance. Techniques like cross-validation are used to avoid
overfitting.
Working and Connectivity
The overall system workflow is outlined below:
4.1 Frontend

• A web-based or desktop interface where the sales team or management

can input variables such as material costs, market trends, and product
details. This interface can be built using frameworks like Flask or
Django.
4.2 Backend

• The machine learning model resides on a backend server or cloud

infrastructure. Once the input data is provided, the model processes it and
returns the predicted price.
4.3 Data Flow

• The system fetches data from the database, processes it through the
trained machine learning model, and then returns predicted prices. The
results are stored back in the database and displayed on the frontend.
4.4 Architecture

• Data Layer: Contains the database with historical sales data, market
trends, etc.
• Application Layer: Consists of the machine learning model and
processing logic.
• Presentation Layer: The user interface where users can interact with the
model and view predictions.

Features of the System

• Automated Price Prediction: Predicts prices of furniture based on input
features such as material cost, product dimensions, and market conditions.
• User-friendly Interface: A simple interface for inputting product details
and viewing the predicted prices.
• Historical Data Analysis: Provides insights into historical price trends
and sales performance.
 • Scalability: Can be easily scaled to accommodate a large number of
product categories and pricing variables.
• Integration with External Data Sources: Can be integrated with APIs to
fetch real-time material costs and competitor pricing information.

Code Implementation
Below is an example of a simple Linear Regression model implemented in
Python using Scikit-learn:

import pandas as pd
from sklearn.model_selection import train_test_split
from sklearn.linear_model import LinearRegression
from sklearn.metrics import mean_absolute_error, mean_squared_error

# Load the dataset

data = pd.read_csv('furniture_sales_data.csv')

# Feature selection
X = data[['material_cost', 'demand_forecast', 'market_trend']]
y = data['price']

# Split the dataset into training and testing sets

X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2,
random_state=42)

# Initialize the Linear Regression model

model = LinearRegression()

# Train the model

model.fit(X_train, y_train)

# Make predictions
y_pred = model.predict(X_test)
# Evaluate the model
mae = mean_absolute_error(y_test, y_pred)
mse = mean_squared_error(y_test, y_pred)

# Output the evaluation results

print(f"Mean Absolute Error: {mae}")
print(f"Mean Squared Error: {mse}")

Test Cases
Test Resul
Input Data Expected Output
Case t

Predicted price close to actual

1 Material Cost: $500, Demand: High, Trend: Up Pass
data

2 Material Cost: $200, Demand: Low, Trend: Down Lower predicted price Pass

Material Cost: $700, Demand: Medium, Trend:

3 Mid-range predicted price Pass
Stable

Advantages
• Accuracy: Machine learning models are more accurate than manual or
heuristic-based price setting.
• Automation: Reduces the need for manual price calculation, saving time
and reducing errors.
• Real-time Price Updates: Enables dynamic pricing adjustments based on
changing market trends and material costs.
• Scalability: The system can handle large datasets, making it scalable for
companies with a vast range of furniture products.
Limitations
• Data Dependency: The accuracy of the model is highly dependent on the
quality and quantity of the data available.
• Complexity: For new product lines or untested markets, the model may
not perform well due to the lack of relevant data.
• External Factors: The model may not account for sudden, unpredictable
factors like supply chain disruptions or economic crises.

Future Scope
• Integration with Real-time Data: Incorporate real-time market data,
such as competitor pricing and material costs, to enable dynamic and up-
to-date pricing.
• Expansion to New Product Categories: Expand the model to predict
prices for new categories or custom-designed furniture.
• Enhanced Predictive Models: Use advanced algorithms such as
XGBoost or Deep Learning for improved accuracy and flexibility in
pricing predictions.
• Mobile Application: Develop a mobile app for real-time access to
predicted prices for the sales team and management.
Conclusion
The implementation of machine learning for price prediction in a large furniture
company provides significant benefits in terms of efficiency, accuracy, and
scalability. By leveraging historical data and market trends, the system
automates the pricing process, reducing manual errors and allowing the
company to stay competitive in a fast-changing market. Though there are
challenges, such as dependency on data quality and external factors, the future
scope of the project offers opportunities to enhance the system further, making
it a vital tool in the company’s operations.

References
[1]. N. Shinde and K. Gawande, “Valuation of House Prices Using Predictive
Techniques,” Int. J. Adv. Electron. Comput. Sci. ISSN2393-2835, vol.
Volume-5, no. 6, pp. 34–40, 2018.
[2]. S. Borde, A. Rane, G. Shende, and S. Shetty, “Real Estate Investment
Advising Using Machine Learning,” Int. Res. J. Eng.
Technol., vol. 4, no. March, pp. 1821–1825, 2017
[3]. M. Bohanec, M. Kljajić Borštnar, and M. Robnik-Šikonja, “Explaining
Machine Learning Models In Sales Predictions,” Expert
Syst. Appl., vol. 71, no. April,pp. 416–428, 2017, doi: 10.1016/
j.eswa.2016.11.010.
[4] Cen Z,Wang J (2019) Crude oil price prediction model with long short term
memory deep learning based on prior knowledge data transfer. Energy
169:160–171
[5] Lago J, De Ridder F, Vrancx P, De Schutter B (2018) Fore-casting day-
ahead electricity prices in europe:the importance of
considering market integration. Appl Energy 211:890–903
[6] Shastri M, Roy S, Mittal M (2019) Stock price prediction using artificial
neural model: an application of big data
[7] Alameer Z, Fathalla A, Li K, Ye H, Jianhua Z (2020) Multistep- ahead
forecasting of coal prices using a hybrid deep learning model. Resour Policy
65:101588
[8] Law S, Paige B, Russell C (2018) Take a look around: using street view and
satellite images to estimate house prices. arXiv preprint arXiv:1807.07155