Battle Of The Portfolio Optimization Methods 

Hi Miki, not really... I didn't have suitable short strategis :(.

Sounds intresting, deffenetly you should try it. Portfolio optimization methods... It is all about diversification. But if you choose a good assets, you can allocate them randomly and you will get a good profit :). We need to predict futere, but it is difficult :)

Anyone teaching you that without outlining how marginal these "modern improvements are" doesn't understand the nature of stock trading. It is very very hard to outperform random, and any slight reproduce-able improvement is already hugely significant. Your ML approach sounds very interesting though, i wish you much luck. I do believe that ML algorithms such as AE or LSTM-VAE or even different types of GANs can help identify patterns that experts would also use to base their investments on. So if you can learn to identify what successful brokers look at as important features and derive a adequate strategy you should be able to make a reliable profit. Assuming investors aren't making 100% of their profit off of market manipulation and inside trading. xD

Hi, hmm... normally most of the software have support documentation with examples..

Hi. :) one possible way is just to run it as additional strategy. What kind of benchmark you are going to use? I have some improved code (similar what I used in the video), I'm still using it for test some portfolio based investment idea.. P.s. :) maybe it can be a good material for the next video

@@CloseToAlgoTrading I am using it for crypto, but realize SPY as benchmark maybe doesn't make sense. Haha 🙊

@@gbh9152 ok.. but it should not be an issue, if I'm still right remember the code, you have two dataframe one for assets and another for benchmark.. It could be whatever you want. BTC for example. However this code is not very sutable if you are going to use something like equally weighted portfolio as benchmark. The simplest way might be run your benchmark and then store the result and compare it with your specific strategy. Or :)) not to use this code directly..

@@CloseToAlgoTrading I see, thanks for sharing. Awesome content! Keep it going. 🙏

Hi.. Hm.. All of them...

Thanks

Hi. Actually I do, but in a slightly different field. 😎 Send me an email with details and your offer...

What will be a good Sharpe for you? It could be for example because of volatility. Momentum strategies are often volatile. This increased risk can lead to a lower Sharpe ratio if the additional returns do not compensate for the additional risk.

All code and files in github. Please check the link in description. Or something is missing there?

Yes ofc, my bad, thanks for the fast reply

Hi, sorry, I don't know. :( I'm not using R.

Hi.. Hmm.. It might be that you use newer version of tensorflow.. Maybe 2.1 or 2.3 will solve issue

ChatGPT ;) Sure, here's some sample code for portfolio optimization using Omega ratio in Python: import numpy as np from scipy.optimize import minimize def omega_ratio(weights, returns, target=0): """Calculate Omega Ratio for given portfolio weights""" portfolio_return = np.sum(returns * weights) downside_returns = np.minimum(target - returns, 0) downside_risk = np.sqrt(np.mean(np.square(downside_returns))) if downside_risk == 0: return np.inf omega = portfolio_return / downside_risk return omega def optimize_portfolio(returns, target=0): """Optimize portfolio using Omega ratio""" num_assets = returns.shape[1] initial_weights = np.repeat(1/num_assets, num_assets) bounds = [(0, 1) for i in range(num_assets)] constraints = ({'type': 'eq', 'fun': lambda x: np.sum(x) - 1}) result = minimize(lambda x: -omega_ratio(x, returns, target), initial_weights, method='SLSQP', bounds=bounds, constraints=constraints) return result.x # Example usage returns = np.random.normal(0, 0.05, size=(100, 3)) weights = optimize_portfolio(returns) print('Portfolio weights:', weights) print('Omega ratio:', omega_ratio(weights, returns)) In this code, the omega_ratio function calculates the Omega Ratio for a given portfolio using the formula: Omega = Portfolio Return / Downside Risk The optimize_portfolio function uses the minimize function from the scipy.optimize library to find the portfolio weights that maximize the Omega Ratio. The function takes in the historical returns of several assets and an optional target return value for downside returns. It returns an array of optimal portfolio weights for the given asset returns. You can use this code by calling the optimize_portfolio function with a 2D NumPy array of asset returns, where each row represents a different time period and each column represents a different asset. The function returns an array of portfolio weights that can be used to construct the optimal portfolio.

Наших всегда слышно :)