This repository contains the solution for the AI Amplify 24 Hour Hackathon by Team United, comprising members Devanshu, Shanthanu, Yash, Ayush, and Musab. We are proud to announce that our team secured the 2nd Runner-Up Position in this highly competitive event.
- Date: 22 July, 2023
- Duration: 24 hours
- Location: 1Finance,Goregaon
Difficulty: HARD
Problem Statement Description: Problem Description: Your task is to develop a Convolution Neural Network (CNN)-based AI system that can accurately recognize and classify different types of flowers – specifically, rose, chamomile, dandelion, sunflower, and tulip. The system should be able to process images of these flowers and accurately classify them into the correct type. Input: Images of five types of flowers provided by the user. These images could vary in terms of flower size, image resolution, and flower orientation.
Output: The type of flower that the input image represents. The system should analyze the input image, recognize the flower, and output the corresponding flower type.
Rules: Participants can use existing CNN architectures to develop the system. However, they should justify their choice of architecture and explain how it contributes to the effectiveness of the system. They should demonstrate the system's ability to accurately recognize a wide variety of flower images.
Difficulty: MODERATE
Problem Statement Description: Problem Description: Your task is to develop an AI-based system that can predict next month's sales for various geographies, distributors, and retailers based on provided sales data. The system should be able to analyze past sales data, understand patterns, account for seasonal trends, and make accurate demand forecasts. The accuracy of the forecast is critical as it can significantly impact the supply chain and business operations. Input: Historical sales data of different geographies, distributors, and retailers. This data could include information such as the number of units sold, revenue generated, time of year, promotional activities, and more.
Output: A forecast of next month's sales for different geographies, distributors, and retailers. The forecast should include a predicted number of units sold and the associated revenue.
Rules: Participants are allowed to use existing forecasting algorithms to develop the system. However, they should justify their choice of tools and explain how these tools contribute to the effectiveness of the system. The participants should be able to demonstrate the accuracy of their forecasting system and its robustness to handle diverse datasets.
Reference: The reference provides a step-by-step guide on how to implement an ARIMA model for inventory demand forecasting in Python. Participants can use this tutorial to understand the problem, learn a potential solution approach, and get a starting point for their own code. https://www.geeksforgeeks.org/inventory-demand-forecasting-using-machine-learning-python/
Difficulty: EASY
Problem Statement Description: Problem Description: Your task is to develop a Recurrent Neural Network (RNN)-based AI system that can analyze product reviews on Amazon and predict the rating based on the review text. The system should be able to understand the sentiment and content of the review and estimate the rating that corresponds to the reviewer's opinion. Input: Amazon product reviews dataset. The dataset would include the review text and possibly other metadata about the review.
Output: Predicted product rating based on the review text. The system should be able to estimate the rating (usually on a 1-5 scale) that corresponds to the sentiment and content of the review.
Rules: Participants can use existing RNN architectures or other suitable machine learning models to develop the system. However, they should justify their choice of models and explain how these contribute to the effectiveness of the system. They should demonstrate the system's ability to accurately predict product ratings based on a wide variety of Amazon reviews.
Reference: The reference is a GeeksforGeeks tutorial demonstrating how to use a Recurrent Neural Network (RNN) for sentiment analysis on Amazon product reviews. Participants can use this tutorial to understand the concept, learn how to implement an RNN for sentiment analysis, and use the provided code as a starting point for their own solution. https://www.geeksforgeeks.org/amazon-product-review-sentiment-analysis-using-rnn/
We won the 2nd Runner-Up Position in this competition.
We would like to thank the event organizers for hosting the event and for keeping the problem statements interesting and engaging! 🚀