Loading project…
The same set-up as project 01, applied to Netflix. The day's open, high, low and volume went in, the same day's close came out, and the rows were shuffled before an 80/20 split.
Original tutorial by Aman Kharwal: Netflix Stock Price Prediction (opens in a new tab)
| Output | Value | Cell # |
|---|---|---|
| Training MSE after epoch 30 | 372.61 |
Cell # counts every cell from the top of the notebook, Markdown included, starting at 1.
A training loss of 372.6 is in squared dollars, so it is not comparable with project 01. The test rows were never scored.
The final cell asked the model for one prediction and failed with NameError because numpy had not been imported in this notebook.
name 'np' is not definedDaily NFLX prices from Yahoo Finance through yfinance, 2009 to 2022.
01 · 05 · 09 · 28 · 32
Conclusion flipsOn a time-ordered split the Apple LSTM misses the close by 38.7% on average, against 1.7% for 'tomorrow equals today'. The linear model in notebook 05 also loses to 'no change' (0.50% versus 0.36%).
Read the audit