This project will understand the customer purchase behaviour (specifically, purchase amount) against various products of different categories. They have shared purchase summary of various customers for selected high volume products from last month.
Banknote analysis refers to the examination of paper currency to determine its legitimacy and identify potential counterfeits. In this python project, I am trying to build a Classification Machine Learning models to predict banknotes are genuine or forged.
SpaceX re-uses the Stage 1 boosters of Falcon 9 rockets. This project is an analysis for successful Stage 1 landing prediction. I have used SpaceX API and Webscraping for data collection, SQlite database for data storage.
Netflix Recommender system is one of the best recommender systems in the world. In this project I've used movies and rating datasets and NLTK toolkit to build this popularity based recommender system.
Netflix Recommender system is one of the best recommender systems in the world. In this project I've used movies datasets and Cosine similarity to build this content based recommender system.
In this project I have built a mechanism to collect information about every book in the website, scraping through pagination. This project is associated with https://books.toscrape.com/ website which is specially design for training web scraping.
A mood classification is a type of machine learning task that is used to recognize human moods or emotions. In this project I have implemented a CNN model for recognizing smiling or not smiling humans using Tensorflow Keras Sequential API.
Sing language is a visual-gestural language used by deaf and hard-to-hearing individuals to convey imformation, thoughts and emotions. In this project I have implemented a CNN model for recognizing sign language digits 0 to 5 using Keras Functional API.
Very deep neural networks suffer from a problem called vanishing/exploding gradients. Deep Residual Learning for Image Recognition resolves theis issue. A Residual Network, also known as ResNet, is a type of deep learning network architecture that introduces the concept of residual learning.
Handwritten digits recognition in deep learning involves training a neural network to accurately classify images of handwritten digits into their corresponding numerical values (0-9). Accurate handwritten digit recognition has broad applications in streamlining processes, improving accuracy, and enabling automation across various industries and fields.
Transfer Learning in Neural Network is a technique used in machine learning where knowledge gained from training one model (source domain) is transferred and applied to a different but related model (target domain). This involves taking a pre-trained model developed for one task and fine-tuned or using its learned features to solve another related task.
LSTM (Long Short-Term Memory) is a type of Recurrent Neural Network (RNN) architecture designed to efficiently capture and utilize long-term dependencies in sequential data. In this project, I will be developping an LSTM model to predict future CO2 emission. CO2 emission data is a Timeseries dataset, which fits for sequential model perfectly.
Countries vary significantly in their CO2 emission levels due to differences in industrialization, energy consumption, transportation, and policies regarding environmental regulations. Some nations, particularly highly industrialized ones like China, the United States, India, and the European Union countries, tend to produce higher CO2 emissions.
A heart attack, medically known as a myocardial infarction, is a serious and potentially life-threatening event that occurs when the blood flow to a part of the heart is blocked. This blockage is often the result of a clot obstructing the coronary arteries, which supply oxygen-rich blood to the heart muscle. When the heart muscle doesn't receive enough oxygen, it can become damaged or start to die, leading to a heart attack.
Diabetes is a chronic health condition characterized by high levels of sugar (glucose) in the blood. It occurs when the body either doesn't produce enough insulin (a hormone that regulates blood sugar) or doesn't effectively use the insulin it produces. This results in an imbalance that can lead to various complications affecting multiple organ systems.
Image captioning in deep learning combines computer vision and natural language processing to generate descriptive text for images. By leveraging neural networks, it automatically generates concise and informative captions that accurately describe the content of an image, enabling applications such as accessibility tools for the visually impaired, content understanding for search engines, and enhancing user experiences in social media and e-commerce platforms.
MedBot was a prototype for remotely controlable robot, designed for COVID wards in hostiptals. This project was one of my volunteer projects which I was participated in 2020. In this project I was one of the main contributors in the team. I developed an Android mobile application as the remote controller application.
Tharanga's Portfolio