Pinned Repositories
CNN_Dog_Breed_Project
A Convolutional Neutral Network built using Python Jupiter Notebook For Udacity Machine Learning Engineer Micro-degree
Corporate-Finance-Projects
In-class projects completed for Corporate Finance Course
Empirical-Method-in-Finance
Winter 2020 Course description: Econometric and statistical techniques commonly used in quantitative finance. Use of estimation application software in exercises to estimate volatility, correlations, stability, regressions, and statistical inference using financial time series. Topic 1: Time series properties of stock market returns and prices Class intro: Forecasting and Finance The random walk hypothesis Stationarity Time-varying volatility and General Least Squares Robust standard errors and OLS Topic 2: Time-dependence and predictability ARMA models The likelihood function, exact and conditional likelihood estimation Predictive regressions, autocorrelation robust standard errors The Campbell-Shiller decomposition Present value restrictions Multivariate analysis: Vector Autoregression (VAR) models, the Kalman Filter Topic 3: Heteroscedasticity Time-varying volatility in the data Realized Variance ARCH and GARCH models, application to Value-at-Risk Topic 4: Time series properties of the cross-section of stock returns Single- and multifactor models Economic factors: Models and data exploration Statistical factors: Principal Components Analysis Fama-MacBeth regressions and characteristics-based factors
Fixed-income-Projects
Basic bond valuation in HW1 and simple strip-principal arbitrage trading algorithm implemented in HW2
ML-and-Data-Science
Machine Learning and Data Mining: Regression [Linear (Selection and Shrinkage, Dimension Reduction, Beyond Linearity) & Non-Linear Regressions (Logistics, K-NN, Trees)], Cross Validation (LOOCV, K-Folds, Bias vs. Variance), Classification (LDA, QDA, K-NN, Logistic, Tree, SVM), Clustering (PCA, K-Means, Hierarchical) This course will provide an introduction to main topics in data mining / statistical learning, including: statistical foundations, data visualization, classification, regression, clustering. Emphasis will be on statistical learning methodology and the models, intuition, and assumptions behind it, as well as applications to real-world problems. You may find my final project in the stats 415 project folder. Project Summary Implemented all the classifier learned throughout the semester to predict obesity rates in America as classified through the BMI with the best classifier as 7-fold KNN and prediction accuracy of 81.54% Analyzed model selection methods to provide the most optimal model and finding the best predictors; concluded that BMI can be non-parametrically predicted based off income, eating habits, exercise habits and shopping habits
Risk-Parity-and-Minimal-Variance-Portfolio-based-on-a-Regularized-Estimate-of-Variance-Covariance-Ma
Estimated the Covariance Matrix with the LedoitWolf and Diagonalization shrinkage methods, significantly reducing the out-of-sample estimating errors Constructed a risk parity and a minimum-variance portfolio using Convex optimization and nested clustered optimization algorithm Backtested the risk parity and minimum-variance portfolios with monthly rebalancing and 49 industry portfolios as asset universe, improving the Sharpe Ratio from 0.23 (no shrinkage) to 0.63
SimpleDemo_AlgoTrading
A Simple demo on how to build a trading algorithm
Statistical-Arbitrage-Hedge-Fund-Strategy-
The goal is to run a backtest simulation using some classical alphas. The backtest is intended to be realistic due to the inclusion of transaction costs.
YiTao-Hu
I am a MFE (Master of Financial Engineering) student at UCLA Anderson School of Management, pursuing a career in quantitative finance or data science
yitaohu88's Repositories
yitaohu88/Empirical-Method-in-Finance
Winter 2020 Course description: Econometric and statistical techniques commonly used in quantitative finance. Use of estimation application software in exercises to estimate volatility, correlations, stability, regressions, and statistical inference using financial time series. Topic 1: Time series properties of stock market returns and prices Class intro: Forecasting and Finance The random walk hypothesis Stationarity Time-varying volatility and General Least Squares Robust standard errors and OLS Topic 2: Time-dependence and predictability ARMA models The likelihood function, exact and conditional likelihood estimation Predictive regressions, autocorrelation robust standard errors The Campbell-Shiller decomposition Present value restrictions Multivariate analysis: Vector Autoregression (VAR) models, the Kalman Filter Topic 3: Heteroscedasticity Time-varying volatility in the data Realized Variance ARCH and GARCH models, application to Value-at-Risk Topic 4: Time series properties of the cross-section of stock returns Single- and multifactor models Economic factors: Models and data exploration Statistical factors: Principal Components Analysis Fama-MacBeth regressions and characteristics-based factors
yitaohu88/Risk-Parity-and-Minimal-Variance-Portfolio-based-on-a-Regularized-Estimate-of-Variance-Covariance-Ma
Estimated the Covariance Matrix with the LedoitWolf and Diagonalization shrinkage methods, significantly reducing the out-of-sample estimating errors Constructed a risk parity and a minimum-variance portfolio using Convex optimization and nested clustered optimization algorithm Backtested the risk parity and minimum-variance portfolios with monthly rebalancing and 49 industry portfolios as asset universe, improving the Sharpe Ratio from 0.23 (no shrinkage) to 0.63
yitaohu88/Statistical-Arbitrage-Hedge-Fund-Strategy-
The goal is to run a backtest simulation using some classical alphas. The backtest is intended to be realistic due to the inclusion of transaction costs.
yitaohu88/CNN_Dog_Breed_Project
A Convolutional Neutral Network built using Python Jupiter Notebook For Udacity Machine Learning Engineer Micro-degree
yitaohu88/Corporate-Finance-Projects
In-class projects completed for Corporate Finance Course
yitaohu88/Fixed-income-Projects
Basic bond valuation in HW1 and simple strip-principal arbitrage trading algorithm implemented in HW2
yitaohu88/ML-and-Data-Science
Machine Learning and Data Mining: Regression [Linear (Selection and Shrinkage, Dimension Reduction, Beyond Linearity) & Non-Linear Regressions (Logistics, K-NN, Trees)], Cross Validation (LOOCV, K-Folds, Bias vs. Variance), Classification (LDA, QDA, K-NN, Logistic, Tree, SVM), Clustering (PCA, K-Means, Hierarchical) This course will provide an introduction to main topics in data mining / statistical learning, including: statistical foundations, data visualization, classification, regression, clustering. Emphasis will be on statistical learning methodology and the models, intuition, and assumptions behind it, as well as applications to real-world problems. You may find my final project in the stats 415 project folder. Project Summary Implemented all the classifier learned throughout the semester to predict obesity rates in America as classified through the BMI with the best classifier as 7-fold KNN and prediction accuracy of 81.54% Analyzed model selection methods to provide the most optimal model and finding the best predictors; concluded that BMI can be non-parametrically predicted based off income, eating habits, exercise habits and shopping habits
yitaohu88/SimpleDemo_AlgoTrading
A Simple demo on how to build a trading algorithm
yitaohu88/YiTao-Hu
I am a MFE (Master of Financial Engineering) student at UCLA Anderson School of Management, pursuing a career in quantitative finance or data science