Been working on the consumer discretionary portfolio since last week. (XLY) Made a few tweaks to it.
In this model ensemble(mix of two models), I try and add to the top 10 when available. Today there is 17, I currently own 9 of the top 13 from the past few days. I have orders in on the remaining 4 in the top 10 that I don't own (green check marks).
The models are looking for 3% moves higher over the next 4-5 days. They use 34 and 42 features/indicators to make predictions and have been trained on 5 and 7 years of data.
The second Image is how this portfolio has done in the past 4 months.
Top: Avg Confidence of the Picks each day (since this model ensemble only outputs high confidence picks, it doesn't fluctuate that much. ) Bar Chart shows how those stocks did on average over the model target (next 4 or 5 days)
Middle : # of Picks per day. When it sees opportunity it suggest stocks, when it doesn't, it goes quiet
Bottom : Equity Curve (teal) vs the Benchmark(equal weight consumer discretionary) . This only assumes 1% per position. Since this is an ensemble of 4-5 day models. Max investment would be 50% at any given time. (Buy 10 stocks at 1% per day (10%) for 5 days, rolling out and in a new 10% everyday)