WONGA : Smart Personal Finance Manager

Group Details

Student ID	Student Name	Usernames	Research Component
IT19974910	Hapugala HAVV	VenuraHapugala	Develop an effective way to automatedly classify expenses to correct expense categories, Give user specific expense insights for each expense category with collected and analyzed expense data
IT19972176	Jayawardana GVHD	HansakaDilshanJayawardana	Automatically read and extract expenses from bills, and SMS using image processing and NLP
IT19961422	Uyanahewa MIR	MadhaviImashi	Develop a mechanism to extract information from the user’s digital calendar and predict expenses for future events
IT19975764	Bandara MBDN	dilendran	Develop an effective way to plan next month’s budget by arranging a personalized automatic solution for the users and assisting them by providing valuable information about their current financial capabilities

Main Objective

In this research, we are endeavoring to develop a smart yet engaging personal finance management mobile application using advanced features to help users improve the efficiency of money management easily and achieve financial success without the need of having any comprehensive knowledge about personal finance management

Objectives

• Automate expense categorization and visual representation to help user understand personal cash flow easily
• Smart expense tracking mechanism to reduce manual data entry
• Predict expenses for upcoming events planned in personal calendars
• Personalized automatic budget planner and chatbot assistant when making financial decisions

Main Research Questions

• How to track daily expenses easily without entering manually
• How to recognize the category of a given expense item without asking the user
• How to determine the efficiency level of user’s money management
• How to predict expenses for future events by using calendar events
• How to implement a system to automatically generate a budget plan

Individual Research Questions

Student ID	Student Name	Questions
IT19974910	Hapugala HAVV	How to analyze the efficiency of user to give category wise user specific expense handling insights, How to classify expenses automatedly to correct expense categories, How to analyze the accuracy of correct expense classification
IT19972176	Jayawardana GVHD	How to track expenditures without enter manually, How to detect bank SMS from an app, How to track bills via snaps without app crashing
IT19961422	Uyanahewa MIR	How to track upcoming events that are already planned by a user, How to predict expenses for upcoming events, How to prevent paying late fees for bills and payments
IT19975764	Bandara MBDN	How to predict user’s future income based on their spending behaviors, How to provide an engaging way to assist users with their financial situations

Individual Objectives

Student ID	Student Name	Objectives
IT19974910	Hapugala HAVV	Creating a proper expense dataset from expenses that were obtained, Choosing and developing a text classification model to classify expenditures into the appropriate categories, Utilizing preprocessed dataset to train the machine learning model to correctly classify expenses, Demonstrate user-based expense behavior insights, Analyze overall app users’ expense data to give user-based insights
IT19972176	Jayawardana GVHD	Data Preprocessing to train, Select appropriate machine learning models, Text extraction, value identification, record transaction details from bills, Detect, extract, and record transaction details from SMS, Develop a mobile application with proper UI/UX
IT19961422	Uyanahewa MIR	Extracting calendar data of users in the required format, Analyze event data written in natural language, Filter future events that require spending money, Predicting expense amounts for the above-identified events, Allow setting reminders for payment dues
IT19975764	Bandara MBDN	Collect and preprocess expense, income data to create appropriate datasets, Select appropriate machine learning models, Train the selected machine learning model using preprocessed datasets, Implement an automatic budget planner using expense predictions, Implement an intelligent chatbot to assist users, Develop a mobile application with proper UI/UX

Technologies

• Languages - Python (3.10), JavaScript
• Frameworks - React-Native, Flask Server
• Database - Firebase
• IDEs - PyCharm IDEA, VS Code
• Integrate Technology Service - GitHub, GitLab
• Hosting - PuTTY, WinSCP, AWS, FastAPI
• Other Technologies - Google Colab, Jupyter Notebook, Tensorflow, Keras, OpenCV, pytesseract, Regex, Numpy, Google App Script, Google API, Fernet, Expo

Algorithms

• Naïve Bayes Multi-Classification Algorithm
• Feedforward Neural Networks (ANN)
• CNN
• Inception V3
• DistilBERT NLP Model
• Decision Tree Regressor
• Gradient Boosting Regressor
• LSTM
• ARIMA
• Text-davinci-003

Techniques

• Data Preprocessing
• Machine Learning Techniques
• OCR
• Data Collection
• Data Pre-processing
• Data Visualization
• Data Encryption
• Transfer Learning
• Keyword Extraction
• Feature Selection
• Hyperparameter Tuning
• Security Permission Handling