Atique — Machine Learning freelancer in Berlin

\u003E Goal The goal for this project was first, to analyze the behavior of GDP on Stock prices and as per client requirements, I used a lot of statistical metrics to explain the concurrent and predictive behaviors and lag it time which it takes stock prices to be effected by GDP. And after this analysis, identification of best times to sell the stocks with knowledge of Hugh risk and high benefit times and low risk, low benefit times for a particular stock using Markov Hidden Regime Detection. \u003E Contribution •\tProposed the summarization of data on a scale which will best explain w.r.t. GDP. i.e. Quarterly and on six months. •\tIdentification of stocks which behave very weirdly and don’t show a consistent behavior. •\tDetection of anomalies in dataset of each stock using inter quartile range criteria. •\tIdentified the fraction difference in the price of each stock w.r.t. GDP. •\tUsed Correlation and p test(significance test) to summarize the behavior of stocks vs GDP •\tIdentified the concurrent and predictive behavior. •\tUsed R-square(variance explained) matrix to identify of each stock w.r.t. GDP to see if we can be able to efficiently and accurately predict the behavior. •\tAnalysis of Cointegration behavior •\tImplementation of model to find the Hidden Markov Regimes. \u003E Results •\tMerged and pruned the data in such a way that it was easier to see the effect and explain the behavior of GDP on stocks. •\tCreated Model for Identification of Hidden Markov Regimes was able to accurately predict such regimes on historical data and then identified regimes were verified to be true as model suggested.