def test_mean_weighted(data):
x = PanelData(data.x)
w = PanelData(data.w)
missing = x.isnull | w.isnull
x.drop(missing)
w.drop(missing)
entity_mean = x.mean("entity", weights=w)
c = x.index.levels[0][get_codes(x.index)[0]]
d = get_dummies(Categorical(c, ordered=True))
d = d[entity_mean.index]
d = d.values
root_w = np.sqrt(w.values2d)
wx = root_w * x.values2d
wd = d * root_w
mu = lstsq(wd, wx, rcond=None)[0]
assert_allclose(entity_mean, mu)
time_mean = x.mean("time", weights=w)
c = x.index.levels[1][get_codes(x.index)[1]]
d = get_dummies(Categorical(c, ordered=True))
d = d[list(time_mean.index)]
d = d.values
root_w = np.sqrt(w.values2d)
wx = root_w * x.values2d
wd = d * root_w
mu = pinv(wd) @ wx
assert_allclose(time_mean, mu)
def test_mean_weighted(data):
x = PanelData(data.x)
w = PanelData(data.w)
missing = x.isnull | w.isnull
x.drop(missing)
w.drop(missing)
entity_mean = x.mean('entity', weights=w)
c = x.index.levels[0][x.index.labels[0]]
d = pd.get_dummies(pd.Categorical(c, ordered=True))
d = d[entity_mean.index]
d = d.values
root_w = np.sqrt(w.values2d)
wx = root_w * x.values2d
wd = d * root_w
mu = np.linalg.lstsq(wd, wx)[0]
assert_allclose(entity_mean, mu)
time_mean = x.mean('time', weights=w)
c = x.index.levels[1][x.index.labels[1]]
d = pd.get_dummies(pd.Categorical(c, ordered=True))
d = d[time_mean.index]
d = d.values
root_w = np.sqrt(w.values2d)
wx = root_w * x.values2d
wd = d * root_w
mu = pinv(wd) @ wx
assert_allclose(time_mean, mu)
def test_mean_weighted(data):
x = PanelData(data.x)
w = PanelData(data.w)
missing = x.isnull | w.isnull
x.drop(missing)
w.drop(missing)
entity_mean = x.mean('entity', weights=w)
c = x.index.levels[0][x.index.labels[0]]
d = pd.get_dummies(pd.Categorical(c, ordered=True))
d = d[entity_mean.index]
d = d.values
root_w = np.sqrt(w.values2d)
wx = root_w * x.values2d
wd = d * root_w
mu = np.linalg.lstsq(wd, wx)[0]
assert_allclose(entity_mean, mu)
time_mean = x.mean('time', weights=w)
c = x.index.levels[1][x.index.labels[1]]
d = pd.get_dummies(pd.Categorical(c, ordered=True))
ilocs = [int(d.columns.get_indexer_for([i])) for i in time_mean.index]
d = d.iloc[:, ilocs]
# TODO: Restore when fixed in pandas
# d = d[time_mean.index]
d = d.values
root_w = np.sqrt(w.values2d)
wx = root_w * x.values2d
wd = d * root_w
mu = pinv(wd) @ wx
assert_allclose(time_mean, mu)
def test_mean_missing(data):
xpd = PanelData(data.x)
xpd.drop(xpd.isnull)
entity_mean = xpd.mean("entity")
expected = xpd.dataframe.groupby(level=0).mean()
expected = expected.loc[xpd.entities]
expected.columns.name = None
assert_frame_equal(entity_mean, expected)
time_mean = xpd.mean("time")
expected = xpd.dataframe.groupby(level=1).mean()
expected = expected.loc[xpd.time]
expected.columns.name = None
assert_frame_equal(time_mean, expected)
dfp.count() == dfp.nobs
`all()`は，列に対して全ての要素が`True`の場合のみ`True`を返すので，これを使い確認できる。
(dfp.count() == dfp.nobs).all()
`( )`はその中を先に評価する，という意味（数学と同じ）。変数が多い場合，`all()`を2回使うと全ての変数に対して評価するので便利である。
(dfp.count() == dfp.nobs).all().all()
`False`なので unbalanced panel data ということが確認できた。
---
変数の観察単位毎の平均の計算
dfp.mean()
---
変数の時間毎の平均の計算
dfp.mean('time')
---
変数の平均からの乖離 $x-\bar{x}$，$\bar{x}$は平均。
dfp.demean()
---
変数の１階階差の計算 $x_t-x_{t-1}$
dfp.first_difference()
