fix strength word

This commit is contained in:
didier
2021-03-27 18:06:07 +00:00
parent b64fcf8359
commit 88b624585c

View File

@@ -533,28 +533,28 @@ def decompose(other_args: List[str], ticker: str, stock: pd.DataFrame):
FtMul = max(0, 1 - np.var(resultMul.resid)) / np.var(
resultMul.trend + resultMul.resid
)
print("Strenght of Trend: %.4f" % FtMul)
print("Strength of Trend: %.4f" % FtMul)
FsMul = max(
0,
1
- np.var(resultMul.resid)
/ np.var(resultMul.seasonal + resultMul.resid),
)
print("Strenght of Seasonality: %.4f" % FsMul)
print("Strength of Seasonality: %.4f" % FsMul)
else:
FtAdd = max(
0,
1 - np.var(resultAdd.resid) / np.var(resultAdd.trend + resultAdd.resid),
)
print("Strenght of Trend: %.4f" % FtAdd)
print("Strength of Trend: %.4f" % FtAdd)
FsAdd = max(
0,
1
- np.var(resultAdd.resid)
/ np.var(resultAdd.seasonal + resultAdd.resid),
)
print("Strenght of Seasonality: %.4f" % FsAdd)
print("Strength of Seasonality: %.4f" % FsAdd)
print("")
except Exception as e: