kevinlu1996 Goto Github PK

credit_fraud_analysis

The purpose of this project is to build an optimal machine learning model that could identify credit card fraud transactions through comparison of different statistical methods and their hyper-parameters tuning. This project works with an imbalanced dataset on credit card transactions. With the nature of an imbalanced dataset, default accuracy would not achieve assessment goal of model performance, thus additional metrics need to be used. I used Recall, Precision, False Positives, False Negatives, and Roc Auc Scores as evaluation metrics. The goal is to optimize machine learning models through hyperparameter tunings to maximize model recall, precision, and roc auc scores and minimize false positives and false negatives.

ethereum-nft-marketplace-boilerplate

gcp-essentials

Example gcloud statements for GCP services

income_classification

kickstarter_visualization

lisa_cnn_image_classification

regression_forecast_ames_housing

This project uses housing price data in Ames, Iowa from Kaggle. The objective is to build a regression model to predict houses with unknown sales price

s-and-p-500-companies

List of companies in the S&P 500 together with associated financials

serverless-gis-with-s3-opendata-and-arcgis

simplehtr

Handwritten Text Recognition (HTR) system implemented with TensorFlow.

webscrapping_classification_transformer

This project uses webscraping to retrieve data from 2 subreddits and identify posts through most frequent words, authors, post lengths, etc, and uses classification model such as Logistic Regression and Voting Classifier with transformer such as CountVectorizer and TFIDFVectorizer to classify posts from each subreddit

weekly

科技爱好者周刊，每周五发布