Research Projects & Events
Reliability of Institutional Broker's Momentum & Oscillator Trading Strategies : An Empirical Study
Research aimed at examining the efficacy of trading strategies promoted by Institutional Brokers to Retail investors.
Python functions developed for 70 popular Trading Strategies used in brokerages such as Fidelity, Interactive Brokers etc.
Empirical Analysis including Portfolio Sorts, PCA, factor regressions were done to study efficacy of trading strategies.
Effect of Firm's Research Expenditure on Innovation Output : Joint work with Finance Ministry, India
Cross-country firm-level analysis to study the effect of research expenditure on innovation output of firms across nations.
Acknowledged for the contribution to "Innovation: Trending Up but Needs Thrust, Especially from the Private Sector" in the Economic Survey of India, 2020
Impact of Demonetization on Tax Aggressiveness Behaviour of Indian Firms : An Econometric Study
Econometric study to determine the impact of 2016 Demonetization on Tax Aggressiveness behavior of Indian firms
Tax aggression measures such as ETR, BTD, Discretionary Permanent Differences, Abnormal BTD were computed.
Difference-in-Difference Analysis on firms with varying exposure to informal economy Post-Demonetisation period .
Regularization based persistence study of financial risk factors & feature importance study
Applied roll-forward PCA for extracting principal risk factors amongst 50 financial features used in Kozak et.al (2020)
Investigated time-varying nature of top principal risk factors to determine the persistence of the risk factors.
Furthermore, applied PCA to construct eigen portfolios and compare the returns against Buy and Hold Strategy .
Guest Speaker : Application of Reinforcement Learning in Financial Markets (Microsoft Learn Series Event on AI)
(Invited as a Guest Speaker at Microsoft Student Learn Event ,hosted by Swati Rajwal ,Microsoft Student Ambassador at NSIT Delhi)
Workshop Highlights
Explained the concepts of Reinforcement Learning algorithms such as Q-value functions, Agent- Environment feedback loop, Deep Q-Network learning Agent, epsilon-greedy algorithm , Bellman Equations etc.
Pulled stock level price data from Yahoo Finance & built a RL trading bot with helper functions, Agent & Environment Classes and DQN Neural Network Architecture on Microsoft Azure using Python .