Department of Computer Science & Engineering
Acharya Institute of Technology
Soladevanahalli, Bengaluru 560 107
Project Phase 1 Synopsis
Presentation on
EstiCheck - A ML Powered Price Estimation and
Reasonability Website for E-Devices
Presented by
Group No Name USN
Dept of CSE
Jyoti Mandal 1AY22CS079
K Nandini 1AY22CS081
29
Kanishka Verma 1AY22CS083
Palak Yadav 1AY22CS122
Guide Name: Mrs. PADMAVATHI C A
Designation: Assistant Professor
1. INTRODUCTION
Motivation:
• Consumers face difficulties in determining the fair price of electronic devices due
to fluctuating market prices.
• This project aims to offer an intelligent system for accurate price estimation and
reasonability.
Dept of CSE
Significance:
• Uses machine learning to analyse pricing patterns and provide reliable price
insights for electronic products.
• Ensures transparency, helping consumers make informed purchasing decisions.
2. LITERATURE SURVEY
Technical Ideas Shortfalls and solution
Name of
Publication / Algorithms provided in the
S.N Paper title the
details discussed in proposed
Authors
the paper methodology
Analysis of centralized
Lacks implementable model;
The Limits of Centralized Arun pricing in dynamic
Dept of CSE 1 Pricing in Online TSE, 2022 Sundararajan et marketplaces; issues
proposed system adds centralized
logic with explainable outputs and
Marketplaces al. with vendor resistance
user validation
and misalignment
Artificial Intelligence Based AI-based price
No real-time data or reasonability
Procurement System for estimation using
scoring; proposed system adds
2 Optimal Price Estimation IJRASET, 2023 Harika K. et al. regression; NLP for
predictive modeling and
and Supplier supplier
centralized validation
Recommendation recommendation
Price Reasonability D. R. Pawar et al. Random Forest, No deployable system or
Determination Model Using electronics domain focus;
3 Machine Learning
IJRECE, 2023 Gradient Boosting used
proposed project builds a web-
for price classification
Techniques based tool for real use
Technical Shortfalls and
Ideas / solution
Publication Name of the
S.N Paper title Algorithms provided in the
details Authors
discussed in the proposed
paper methodology
Review of ML Lacks domain-specific
algorithms (Decision focus; proposed system
Dept of CSE 4 A Survey on Machine Learning
IJSRET, 2022 M. Patel and A. Joshi Trees, SVM, Neural applies techniques to
Techniques for Price Prediction
Networks) for price electronics with real-
prediction world data pipelines
Association for Random Forest best;
Effective E-commerce Price 8 models + Bayesian
5 Prediction Using ML
ACM ASENS, 2024 Computing
optimization
tuning helps more than
Machinery (ACM) sentiment analysis
No implementation;
Theoretical comparison
Centralized Pricing vs. supports our ML-based
of centralized vs.
6 Delegating Pricing Under INFORMS, 2021 Anderson and Prasad
delegated pricing under
centralized pricing
Information Asymmetry framework for objectivity
asymmetric info
and transparency
3. PROBLEM STATEMENT &
OBJECTIVES
Problem Statement:
• Online electronics prices vary across platforms, creating confusion, with no transparent
way to verify if a price is reasonable. Static datasets fail to account for market
dynamics, product specs, location, and competition.
Objectives:
Dept of CSE • The objective is to develop a Centralised ML-based system Website that utilizes
regression models and Random Forest algorithms to predict and validate the prices of
electronic devices, providing users with reliable price assessments.
• The system will process real-time data, considering factors such as product specs,
demand, location, and market competition, to provide accurate and up-to-date price
validation for electronic devices.
4. PROPOSED METHODOLOGY
1.Data Collection:
1.Scrape real-time product listings and historical data from platforms like Amazon,
Flipkart, Yahoo Shopping.
2.Collect features: brand, model, RAM, storage, processor, location, etc.
2.Data Preprocessing:
1.Clean, normalise, and encode technical specifications.
2.Handle missing or inconsistent data.
3.Model Training:
Dept of CSE 1.Use ML algorithms Random Forest and Linear Regression to predict and validate.
2.Train model to predict price based on specifications and market factors.
4.Reasonability Analysis:
1.Compare listed price with model-predicted price.
2.Generate reasonability verdict: Underpriced, Fair, or Overpriced.
5.Web Application Interface:
1.Input form for user to enter product details and price.
2.Backend API to trigger prediction and return output.
3.UI displays predicted price, verdict, and key influencing factors.
5. DELIVERABLES & IMPACT
Deliverables:
1. A fully functional web application for price estimation and validation of electronic
devices.
2. Integration of ML models for accurate price prediction and reasonability checks.
3. Real-time data processing for dynamic price validation based on product
specifications, demand, location, and market competition.
4. Visual dashboard with dynamic charts for price trends and market movements.
5. Product comparison feature to analyse pricing and popularity across platforms.
Dept of CSE 6. Manuals: Comprehensive user and technical documentation for easy navigation and
understanding of the system.
Impact:
• Business Impact: Optimizes pricing strategy to increase profitability and capture
customer interest.
• Operational Impact: Automates pricing and validation, speeding up time-to-market
and improving adaptability.
• Customer Impact: Builds trust with fair pricing and enhances the shopping experience.
• Technical Impact: Offers a reusable and scalable architecture for future ML projects.
6. GROUP MEMBERS INDIVIDUAL RESPONSIBILITIES
Jyoti Mandal K Nandini Kanishka Verma Palak Yadav
•Conduct literature •Build and test • Design the backend • Connect frontend
review and machine learning system architecture with backend API
competitor analysis models (LR, RF) • Create API • Create visualizations
•Design web scraping •Work on demand endpoints for price for price
strategies prediction and price prediction and comparisons and
Dept of CSE •Lead data cleaning reasonability logic validation fairness flags
and preprocessing •Analyze model • Ensure scalability • Conduct user testing
•Design a clean and performance and and modularity and polish the final
intuitive frontend iterate output
• Work with modeling
team to define
features
7. GANTT CHART
Dept of CSE
8.RESULTS
Dept of CSE
Dept of CSE
Dept of CSE
Dept of CSE
8. Discussion on Objective 1
Front end construction and Price Estimation of Electronic Device
Frontend Overview:
• Built using React (with TypeScript) for component-based UI.
• Styled using Tailwind CSS.
• Uses .tsx files for dynamic components and .ts for logic separation.
• Responsive design for mobile and desktop views.
Dept of CSE
Price Estimation:
• Built a machine learning model using Random Forest for price estimation.
• Trained on real e-commerce data.
• Accurately predicts price based on product specifications.
• Integrated with the frontend to show results dynamically.
8. REFERENCES
1. A. Sundararajan, et al., “The Limits of Centralised Pricing in Online
Marketplaces,” IEEE Transactions on Software Engineering (TSE), vol. 48,
no. 3, pp. 601–612, 2022.
2. Harika K., K. S. Srujana, and M. Vijayalakshmi, “Artificial Intelligence
Based Procurement System for Optimal Price Estimation and Supplier
Recommendation,” International Journal for Research in Applied Science and
Engineering Technology (IJRASET), vol. 11, no. 4, pp. 788–794, 2023.
3. D. R. Pawar, A. R. Shaikh, and P. P. Shinde, “Price Reasonability
Determination Model Using Machine Learning Techniques,” International
Dept of CSE
Journal of Recent Engineering and Communication (IJRECE), vol. 11, no. 1,
pp. 45–49, 2023.
4. M. Patel and A. Joshi, “A Survey on Machine Learning Techniques for Price
Prediction,” International Journal of Scientific Research in Engineering and
Technology (IJSRET), vol. 9, no. 5, pp. 134–138, 2022.
5. E. Anderson and A. Prasad, “Centralised Pricing vs. Delegating Pricing Under
Information Asymmetry,” INFORMS Journal on Marketing Science, vol. 40,
no. 2, pp. 220–232, 2021.
