KuanWeiBeCool

Binary-Decimal-Converter

This repository contains the converter I designed that can convert the following:

Build-InfoGAN-From-Scratch

GANs are widely used for synthesizing new data, especially images. However, one drawback of normal GANs is that we have no control over the images GANs produce. For instance, a GAN that is trained to produce fake hand-written digit images may be able to generate very real hand-written digit images, but we have no control over which number it generates. InfoGAN solved this problem: the network can learn to produce images with specific categorical features (such as digits 0 to 9) and continuous features (such as the rotational angle of the digits), in an unsupervised manner. In addition, because the learning is unsupervised, it is able to find the patterns hidden among the images, and generate images that follow these hidden patterns. Sometimes the model can learn very interesting patterns that are beyond your imagination (for example, one of my models learns to transit from number 2 to number 8. You will see it later!). In this notebook, I will introduce how InfoGAN achieves the control of the images being produced, and how to build an InfoGAN from scratch to synthesize feature-specific MNIST hand-written digits.

Choose-best-Sephora-Make-up-Products-With-A-Limited-Budget

This repository contains the script for infinite scroll web page scraping developed by myself, and a practical project which uses this scraping code for collecting product information from Sephora and select best combinations of new make-up within a given budget of 100 dollars.

Connect-Four-Game

This project contains the HTML code and the JavaScript for a web app that allows the users to play "Connect Four"

Data-Project-How-Do-Education-and-Income-Affect-Marriage-and-Divorce-Rates

This project studies the correlations between marriage rate, divorce rate and income, education, to try to understand some conflicts in the perspectives of marriage in between our generation and our parent's generation

Full-Featured-Question-and-Answer-Web-App-Public

Implementing-A-Multi-Layer-Artificial-Neuron-Network-From-Scratch

In this notebook, I share a multi-layer artificial neural network that I developed from scratch with the feature of customizing the number of layers and the number of neurons in each layer. I also include the full math derivations using two simple examples to illustrate the principle of the algorithm. At the end, I also walk through the most important function in my classifier - the "fit" function - and explain step by step what the code does.

Object-Extraction-From-Images

Predict-Electricity-Demand-in-Ontario

This repository contains files and code for the project - "Forecasting Ontario’s Electrical Demand Using Machine Learning" - authored by Kuan Wei, Lucas Crea, Manuel Sage, and Jiarui Xie. We collected the hourly electricity demand in Ontario from the years 2017 to 2020 (https://www.ieso.ca/en/Power-Data/Data-Directory). Other features used for the predictions include time (converted into sine-cosine encoding), temperature (a weighted average temperature across six weather stations in the major population centers across the province: Hamilton, Kitchener, London, Ottawa, Toronto and Windsor, from https://climate.weather.gc.ca), and holiday information. Four machine learning models (RF, FCNN, LSTM, GRU) are used and their model performances are compared.

Weather-Predictions-Using-Deep-Learning

KuanWeiBeCool/Choose-best-Sephora-Make-up-Products-With-A-Limited-Budget

This repository contains the script for infinite scroll web page scraping developed by myself, and a practical project which uses this scraping code for collecting product information from Sephora and select best combinations of new make-up within a given budget of 100 dollars.

KuanWeiBeCool/Data-Project-How-Do-Education-and-Income-Affect-Marriage-and-Divorce-Rates

This project studies the correlations between marriage rate, divorce rate and income, education, to try to understand some conflicts in the perspectives of marriage in between our generation and our parent's generation

KuanWeiBeCool/Chinook-Music-Store-Data-Analysis-Using-SQL

Chinook database is a media-related database that contains data about Chinook Music Store. The database contains 11 different tables: album, artist, customer, employee, genre, invoice, invoice_line, media_type, playlist, playlist_track, and track. In this project I will apply my SQL skills to answer several business-related questions.

KuanWeiBeCool/Kaggle-Titanic-Competition

Titanic: Machine Learning from Disaster is a famous Kaggle competition in which participants are asked to build machine learning models to predict if a passenger will survive or not. It is a great resource for practicing data analysis skills for self-learning students like myself. In this notebook, I present how I handle the dataset and eventually build a model with a testing dataset prediction accuracy of 78.7%.

KuanWeiBeCool/Understand-Transposed-Convolutions-And-Build-Conv2DTranspose-Layer-From-Stratch

In generative adversarial network (GAN), convolutions and transposed convolutions are both heavily involved. While convolutions play an important role in the discriminator, transposed convolutions are the primary building blocks for the generator. The tensorflow API - Keras - has made building GAN a very convenient process. However, sometimes it can be confusing of **what values should be used for the kernel size, strides, and padding to yield the right output shapes.** Setting the right values for the parameters require us to understand how transposed convolutions work. In this notebook, I would like to share some of my personal understandings about transposed convolutions. Throughout the notebook, I will use convolutions as the comparison to better explain transposed convolutions. I will also show you how I implement these understanding to build my own convolutional and transposed convolutional layers, which act as a naive version of the Conv2D and Conv2DTranspose layers from Keras. The notebook consists of three sections: #### 1. What are transposed convolutions? #### 2. How do the parameters (kernel size, strides, and padding) affect transposed convolutions? #### 3. Build my own Conv2D and Conv2DTranspose layers from scratch