Hey There!
I'm Dixon.
I'm a passionate Data Science / AI student from Kuala Lumpur, Malaysia.
Hey There!
Projects
AutoGate AI
AutoGate AI is a state-of-the-art automated gate control system specifically designed for secure and efficient vehicle access management in Malaysia. Leveraging the Ultralytics YOLOv10 model for real-time license plate detection and EasyOCR for accurate text extraction, the system cross-references this data with a centralized MongoDB database to determine access permissions. It features a user-friendly Python Flask web interface, enabling seamless management of resident and visitor records, as well as detailed access logs. Engineered with Non-Maximum Suppression (NMS) for optimal accuracy, AutoGate AI delivers reliable, automated gate control at both entry and exit points, ensuring heightened security and operational efficiency.
Learn MoreReal-Time
Cryptocurrency Prediction Model
The Real-Time Cryptocurrency Prediction Model project, part of my Diploma in Computer Science final year project, seeks to revolutionize cryptocurrency market predictions through advanced machine learning and statistical techniques following the CRISP-DM methodology. My partner and I ensure data quality by collecting information from reputable exchanges like Yahoo Finance and Binance. To enhance user experience, we've developed a user-friendly online web application that delivers accurate predictions, marking a significant stride in the realm of cryptocurrency analysis.
Credit Card Fraud
Detection
Credit Card Fraud Detection is an innovative project focused on creating a robust system for detecting fraudulent credit card transactions within a dataset of over 284,000 records, marked by a significant class imbalance. The project leverages three machine learning algorithms—Logistic Regression, Decision Tree, and Random Forest—to effectively analyze transaction data. To address the challenge of class imbalance, the Synthetic Minority Over-sampling Technique (SMOTE) is employed to generate synthetic samples for the minority class, thereby improving the models’ detection capabilities. By integrating these advanced methodologies, the project aims to deliver a reliable and accurate solution for credit card fraud detection, underscoring the transformative potential of machine learning in enhancing financial security.
Learn MoreBackpack.TF Listings
Backpack.TF Listings is a FastAPI application that streamlines interactions with Backpack.TF, a leading platform for real-time listings of virtual items in Team Fortress 2. This application harnesses websocket connections to deliver live updates and employs the Backpack.TF API to efficiently fetch and store listings data in a MongoDB database. Users can easily access this information through a well-structured RESTful API, making it an indispensable tool for traders and developers in the Team Fortress 2 community. Designed for optimal performance, it can process over 10,000 listings per minute and includes optional user authentication for secure API access.
Learn MoreSteam Inventory Fetcher
Steam Inventory Fetcher is a sophisticated tool designed to optimize the retrieval of Steam user inventories. It employs an advanced smart proxy rotation system that effectively manages high request volumes, significantly reducing the risk of rate limiting. Specifically engineered for seamless integration with WebShare proxies, this application ensures reliable performance and minimizes the likelihood of IP bans. Key features include automatic proxy rotation, a cooldown mechanism for rate-limited proxies, and an efficient caching system to streamline operations. Additionally, the built-in fallback mechanism and a retry limit enhance the tool’s resilience, making it an invaluable resource for developers and traders within the Steam community.
Learn MoreBengaluru
House Price Prediction Model
The Bengaluru House Price Prediction Model utilizes the robust Ordinary Least Squares (OLS) regression analysis to deliver precise and dynamic forecasts for property prices in the rapidly evolving real estate market of Bengaluru. By harnessing a comprehensive dataset that includes intricate details such as property features and location attributes, the model ensures accuracy in capturing the nuanced dynamics inherent to the Bengaluru real estate landscape. This sophisticated approach not only enhances the reliability of our predictions but also equips users, whether buyers, sellers, or investors, with invaluable insights to navigate the intricate and fast-paced real estate environment of Bengaluru with confidence.
Customer
Purchase Online Intention Prediction Model
The Customer Purchase Online Intention Prediction Model is a sales optimization solution employing a Logistic Regression algorithm to predict online customer purchase intentions. By analyzing diverse attributes from e-commerce websites and customer data, such as browsing history and demographic information, the model offers real-time predictions of customer intent. This enables businesses to tailor marketing strategies, provide personalized recommendations, and enhance overall customer experience, ultimately boosting sales by aligning with customer preferences in the dynamic online marketplace.
Sunway
Pyramid
Shopping Mall Parking System
The parking system is designed to efficiently handle parking fees for vehicles. This system incorporates procedures for vehicle check-in and check-out, tracks parking duration, and calculates the corresponding cost based on the time spent in the parking facility. By seamlessly managing these processes, the system ensures a streamlined and accurate approach to parking fee computation, providing users with a convenient and transparent experience for parking transactions.
TF2autobot
Autopricer
The TF2Autobot Autopricer, powered by an advanced Application Programming Interface (API) and utilizing extensive listing data sourced from Backpack.TF, distinguishes itself not only for its precision in generating dependable prices for TF2Autobot but also for its resilient mechanism engineered to proactively prevent and combat price manipulation. Through the integration of sophisticated algorithms and continuous monitoring, it reinforces the integrity of pricing data, serving as a formidable deterrent against manipulative tactics within the Team Fortress 2 trading community.
TF2
Arbitrage Bot
The TF2 Arbitrage Bot stands as an advanced system seamlessly integrated with TF2Automatic, specifically engineered to identify and exploit arbitrage opportunities spanning various Team Fortress 2 websites including popular websites such as Backpack.TF and others. This cutting-edge bot operates by autonomously acquiring items from one user and subsequently selling them to another, strategically leveraging price differentials to generate profit. Fueled by intelligent algorithms, the bot continually scans a diverse array of Team Fortress 2 websites, ensuring swift and strategic arbitrage actions. By streamlining the entire arbitrage process, TF2 Arbitrage Bot provides an efficient and automated solution for adeptly navigating the dynamic TF2 marketplace, optimizing trading opportunities, and ultimately maximizing returns.
python-tf2-utilities
It is a Python package for getting information about Team Fortress 2 items, effects, skins and more. Inspired by TF2autobot's node-tf2-schema and TF2autobot's node-tf2-sku.
Learn More