0% found this document useful (0 votes)
5 views3 pages

OCBC AI Transaction & Sentiment Analysis

OCBC Bank is developing an AI-powered application to clean and analyze credit card transactions and customer feedback using Large Language Models (LLMs) and real-time data visualization. The system features data processing, sentiment analysis, real-time insights, and a user-friendly interface, with a focus on scalability and efficiency. Future enhancements include advanced personalization, multi-language support, and integration with enterprise analytics tools.

Uploaded by

Arham Ali
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
5 views3 pages

OCBC AI Transaction & Sentiment Analysis

OCBC Bank is developing an AI-powered application to clean and analyze credit card transactions and customer feedback using Large Language Models (LLMs) and real-time data visualization. The system features data processing, sentiment analysis, real-time insights, and a user-friendly interface, with a focus on scalability and efficiency. Future enhancements include advanced personalization, multi-language support, and integration with enterprise analytics tools.

Uploaded by

Arham Ali
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd

OCBC Bank - AI-Powered Transaction & Sentiment Analysis System

1. Project Overview
This project aims to develop a web and mobile application that leverages Large
Language Models (LLMs) and real-time data visualization to clean and
analyze OCBC’s credit card transactions and customer feedback.
The solution will:
 Automate cleaning of messy/unstructured transaction data.
 Identify sentiment trends in customer feedback.
 Generate real-time data insights for internal OCBC users (IT & Marketing).
 Provide an intuitive, user-friendly interface for seamless interaction.
 Ensure scalability, accuracy, and efficiency in processing large datasets.

2. System Features & Functionalities


A. Data Processing & Cleaning
 Accepts multiple file formats (CSV, JSON, Excel).
 Validates file type before processing.
 Classifies as either Customer Feedback or Credit Card Transactions.
 Cleans and standardizes messy transaction data using LLM-driven NLP
techniques.
 Detects spam/irrelevant customer feedback and filters out noise.
B. Sentiment Analysis & Real-Time Insights
 Uses DeepSeek / Llama 3.1:3B for high-accuracy analysis.
 Option to switch to Paraphrase-MPNet-Base-V2 for faster but less
accurate insights.
 Generates real-time sentiment tracking with Power BI / Tableau /
[Link].
C. Real-Time Data Visualization
 Live dashboard to track sentiment trends.
 Color-coded indicators for negative spikes (e.g., red for major
complaints).
 Historical data trends for long-term insights.
 Real-time alerting on major sentiment shifts.
D. User Interface
1. Web Application
 Developed in [Link] with a [Link] & FastAPI backend.
 Intuitive interface with a hamburger menu for:
o Customer Feedback Analysis

o Credit Card Transaction Cleaning

 Interactive data filtering & export functionality.


2. Mobile Mini-App (Android Studio)
 Streamlined interface for sentiment tracking on the go.
 Ability to upload and analyze customer feedback via mobile.
E. Backend & Infrastructure
 Database: MongoDB (scalable, handles large unstructured datasets).
 Queue System: Redis / RabbitMQ to manage multiple file uploads.
 Pre-processing pipeline to optimize LLM response time.

3. Technical Architecture
 Frontend: [Link] (Web), Kotlin (Android Mini-App)
 Backend: FastAPI (Python) & [Link] for API integration
 Database: MongoDB for high-speed data retrieval
 ML Model Hosting: Hugging Face / AWS Sagemaker
 Visualization: Power BI / Tableau / [Link]
 Queue Management: Redis / RabbitMQ
 LLM Models: DeepSeek / Llama 3.1:3B + Paraphrase-MPNet-Base-V2

4. Project Phases

Phase Tasks

Sandbox Model training, accuracy testing, data validation

Pre- UI design, backend integration, performance


Production optimization

Full-scale deployment, internal testing, future


Production
enhancements

5. Challenges Faced
1. Accuracy vs. Speed Trade-off
o LLMs can be slow, requiring optimizations.

o Used distilled models for faster processing in some cases.

2. Real-Time Data Processing Bottlenecks


o Optimized pipeline with batch processing & queueing.

3. Model Training Challenges


o Faced data scarcity issues, overcame by generating synthetic
feedback data.
4. Scalability Concerns
o Ensured MongoDB indexing & Redis caching for smooth
performance.
5. User Adoption & UI Design
o Iterated UI based on feedback for better usability.

6. Future Enhancements
 Advanced personalization per OCBC department’s needs.
 Multi-language support for regional teams.
 Integration with OCBC’s enterprise analytics tools.
 Production-scale optimization & load testing.

7. References (APA Style)


 Brown, T. et al. (2020). “Language Models Are Few-Shot Learners.”
NeurIPS.
 Devlin, J., et al. (2018). “BERT: Pre-training of Deep Bidirectional
Transformers for Language Understanding.” ACL.
 Microsoft. (2021). “Paraphrase-MPNet-Base-V2: Sentence Embedding
Model.” Hugging Face Models.

You might also like