I am coding the model for the thesis, and do the data analysis, but i will need

I am coding the model for the thesis, and do the data analysis, but i will need someone to write it.
Topic: Volatility Estimation and Forecasting using Deep Learning Models
Description: The volatility of financial assets is a very important metric in quantitative finance, for a number of reasons. Portfolio allocation, option pricing and risk-management models are dependent on the accuracy of estimation and forecast of the underlying assets’ volatility. Standard volatility forecasting techniques include simple estimators of past realized volatility and more advanced econometric models like the GARCH. A key application of these models is to predict the future realized volatility of an asset using past realizations of returns and volatility. In this project, the student may experiment with using modern machine learning techniques like deep neural networks, convolution all neural networks or recurrent neural networks to forecast future volatility of financial assets. For instance, the models could be trained to forecast realized volatility of a stock using past realizations of returns, trading volumes, microstructure-based measures of that specific stock or from other assets. Another potentially interesting topic of research is the usage of the abovementioned machine learning models to estimate intra-volatility using daily stock data. Training could involve observed intra-day realized volatility computed from TAQ data, and the trained model could be used to estimate intra-day volatility in less recent years or for other assets for which intra-day data is not available. Depending on the specific project devised by the student, the work could be based on historical intra-day prices of equity stocks (TAQ data), on daily stock returns (CRSP data), on intra-day limit order book data for Nasdaq equity stocks, on intra-day transaction-level data for Bitcoin or a combination of the above. Requirements Familiarity with time-series econometrics, in particular the estimation of GARCH models. Experience with deployment of machine learning methods in Python, Matlab, or R.
Literature: Stock and Watson, Introduction to Econometrics (4th Edition) – Chapter 17.5 Abdi and Ranaldo 2017. A simple estimation of bid-ask spreads from daily close, high, and low prices. The Review of Financial Studies, 30(12), pp.4437-4480. Peng et al 2018. The best of two worlds: Forecasting high frequency volatility for cryptocurrencies and traditional currencies with Support Vector Regression

Leave a Reply

Your email address will not be published.