def update(self, x, y_pred, centers, sample_weight=1.0):
self._minimum_separation = self._find_minimum_separation(centers)
self._center_centers = {i: stats.Mean() for i in x}
for i in self._center_centers:
for j in centers:
self._center_centers[i].update(centers[j][i], w=sample_weight)
center_centers = {
i: self._center_centers[i].get()
for i in self._center_centers
}
beta_t = stats.Mean()
for i in centers:
beta_t.update(
utils.math.minkowski_distance(centers[i], center_centers, 2))
self._beta_t = beta_t.get()
try:
self._n_points_by_cluster[y_pred] += 1
except KeyError:
self._n_points_by_cluster[y_pred] = 1
self._n_clusters = len(centers)
return self
def update(self, x, y_pred, centers, sample_weight=1.0):
if not self._initialized:
self._center_all_points = {i: stats.Mean() for i in x}
self._initialized = True
for i in self._center_all_points:
self._center_all_points[i].update(x[i], w=sample_weight)
center_all_points = {
i: self._center_all_points[i].get()
for i in self._center_all_points
}
self._n_points += 1
try:
self._n_points_by_clusters[y_pred] += 1
except KeyError:
self._n_points_by_clusters[y_pred] = 1
for i in centers:
self._squared_distances[i] = utils.math.minkowski_distance(
centers[i], center_all_points, 2)
return self
def update(self, x, y_pred, centers, sample_weight=1.0):
self._furthest_cluster_distance = self._find_furthest_cluster_distance(centers)
if not self._initialized:
self._center_all_points = {i: stats.Mean() for i in x}
self._dim = len(x)
self._initialized = True
for i in self._center_all_points:
self._center_all_points[i].update(x[i], w=sample_weight)
center_all_points = {
i: self._center_all_points[i].get() for i in self._center_all_points
}
distance_point_cluster_center = math.sqrt(
utils.math.minkowski_distance(centers[y_pred], x, 2)
)
distance_point_center = math.sqrt(
utils.math.minkowski_distance(center_all_points, x, 2)
)
self._ssq_points_cluster_centers += distance_point_cluster_center
self._ssq_points_center += distance_point_center
self._n_clusters = len(centers)
# To trace back
self.sample_correction = {
"distance_point_cluster_center": distance_point_cluster_center,
"distance_point_center": distance_point_center,
}
return self
def __init__(
self,
optimizer: optim.Optimizer = None,
loss: optim.losses.Loss = None,
l2=0.0,
initializer: optim.initializers.Initializer = None,
clip_gradient=1e12,
seed=None,
):
super().__init__(seed=seed)
self.optimizer = optim.SGD() if optimizer is None else copy.deepcopy(
optimizer)
self.u_optimizer = (optim.SGD()
if optimizer is None else copy.deepcopy(optimizer))
self.i_optimizer = (optim.SGD()
if optimizer is None else copy.deepcopy(optimizer))
self.loss = optim.losses.Squared() if loss is None else loss
self.l2 = l2
if initializer is None:
initializer = optim.initializers.Zeros()
self.initializer = initializer
self.clip_gradient = clip_gradient
self.global_mean = stats.Mean()
self.u_biases: typing.DefaultDict[
int, optim.initializers.Initializer] = collections.defaultdict(
initializer)
self.i_biases: typing.DefaultDict[
int, optim.initializers.Initializer] = collections.defaultdict(
initializer)
def __init__(
self,
n_factors=10,
bias_optimizer: optim.Optimizer = None,
latent_optimizer: optim.Optimizer = None,
loss: optim.losses.Loss = None,
l2_bias=0.0,
l2_latent=0.0,
weight_initializer: optim.initializers.Initializer = None,
latent_initializer: optim.initializers.Initializer = None,
clip_gradient=1e12,
seed: int = None,
):
self.n_factors = n_factors
self.u_bias_optimizer = (
optim.SGD() if bias_optimizer is None else copy.deepcopy(bias_optimizer)
)
self.i_bias_optimizer = (
optim.SGD() if bias_optimizer is None else copy.deepcopy(bias_optimizer)
)
self.u_latent_optimizer = (
optim.SGD() if latent_optimizer is None else copy.deepcopy(latent_optimizer)
)
self.i_latent_optimizer = (
optim.SGD() if latent_optimizer is None else copy.deepcopy(latent_optimizer)
)
self.loss = optim.losses.Squared() if loss is None else loss
self.l2_bias = l2_bias
self.l2_latent = l2_latent
if weight_initializer is None:
weight_initializer = optim.initializers.Zeros()
self.weight_initializer = weight_initializer
if latent_initializer is None:
latent_initializer = optim.initializers.Normal(sigma=0.1, seed=seed)
self.latent_initializer = latent_initializer
self.clip_gradient = clip_gradient
self.seed = seed
self.global_mean = stats.Mean()
self.u_biases: typing.DefaultDict[
int, optim.initializers.Initializer
] = collections.defaultdict(weight_initializer)
self.i_biases: typing.DefaultDict[
int, optim.initializers.Initializer
] = collections.defaultdict(weight_initializer)
random_latents = functools.partial(
self.latent_initializer, shape=self.n_factors
)
self.u_latents: typing.DefaultDict[
int, optim.initializers.Initializer
] = collections.defaultdict(random_latents)
self.i_latents: typing.DefaultDict[
int, optim.initializers.Initializer
] = collections.defaultdict(random_latents)
def update(self, x, y_pred, centers, sample_weight=1.0):
self._minimum_separation = self._find_minimum_separation(centers)
distance = math.sqrt(utils.math.minkowski_distance(centers[y_pred], x, 2))
if y_pred in self._avg_cp_by_clusters:
self._avg_cp_by_clusters[y_pred].update(distance, w=sample_weight)
else:
self._avg_cp_by_clusters[y_pred] = stats.Mean()
self._avg_cp_by_clusters[y_pred].update(distance, w=sample_weight)
return self
def __init__(self,
x: float,
y=typing.Union[float, utils.VectorDict],
weight: float = 1.0):
self.x_stats = stats.Mean()
self.x_stats.update(x, weight)
self.y_stats: typing.Union[stats.Var, utils.VectorDict]
self._update_estimator: typing.Callable[
[typing.Union[float, utils.VectorDict], float], None]
self.is_single_target = True
self._init_estimator(y)
self._update_estimator(y, weight)
def load_stats():
for _, obj in inspect.getmembers(importlib.import_module("river.stats"), inspect.isclass):
try:
if issubclass(obj, stats.Link):
yield obj(stats.Shift(1), stats.Mean())
continue
sig = inspect.signature(obj)
yield obj(
**{
param.name: param.default if param.default != param.empty else 1
for param in sig.parameters.values()
}
)
except ValueError:
yield obj()
def eval_relations(self, model, dataset):
metrics = collections.OrderedDict({
f"{metric}": stats.Mean()
for metric in ["MRR", "MR", "HITS@1", "HITS@3", "HITS@10"]
})
with torch.no_grad():
metrics = self.compute_score(
model=model,
test_set=self.get_relation_stream(dataset),
metrics=metrics,
device=self.device,
)
return {
f"{name}_relations": round(metric.get(), 4)
for name, metric in metrics.items()
}
def eval(self, model, dataset):
"""Evaluate selected model with the metrics: MRR, MR, HITS@1, HITS@3, HITS@10"""
metrics = collections.OrderedDict({
metric: stats.Mean()
for metric in ["MRR", "MR", "HITS@1", "HITS@3", "HITS@10"]
})
with torch.no_grad():
for test_set in self.get_entity_stream(dataset):
metrics = self.compute_score(model=model,
test_set=test_set,
metrics=metrics,
device=self.device)
return {
name: round(metric.get(), 4)
for name, metric in metrics.items()
}
def update(self, x, y_pred, centers, sample_weight=1.0):
if not self._initialized:
self._center_all_points = {i: stats.Mean() for i in x}
self._initialized = True
for i in self._center_all_points:
self._center_all_points[i].update(x[i], w=sample_weight)
center_all_points = {
i: self._center_all_points[i].get()
for i in self._center_all_points
}
squared_distance_center = utils.math.minkowski_distance(
x, center_all_points, 2)
squared_distance_cluster_center = utils.math.minkowski_distance(
x, centers[y_pred], 2)
self._ssq_point_center += squared_distance_center
self._ssq_point_cluster_centers += squared_distance_cluster_center
return self
def _unit_test_params(cls):
return {"statistic": stats.Mean()}
def __init__(self, seed=None):
super().__init__(seed=seed)
self.variance = stats.Var()
self.mean = stats.Mean()
self.seed = seed
def __init__(self, regressor: base.Regressor, window_size: int = None):
self.regressor = regressor
self.window_size = window_size
self.mean = (stats.Mean() if self.window_size is None else
stats.RollingMean(self.window_size))
def __init__(self):
self.x_m = stats.Mean()
self.g_var = stats.Var()
self.h_var = stats.Var()
self.gh_cov = stats.Cov()
X_y = iter(datasets.Bikes())
# Peer one data
# for x, y in X_y:
# pprint(x)
# print(f'Number of avaliable biles: {y}')
# break
# construct a model pipeline
model = compose.Select('clouds', 'humidity', 'pressure', 'temperature', 'wind')
# To extract feature (bike number by hour)
model += (
get_hour
| feature_extraction.TargetAgg(by=['station', 'hour'], how=stats.Mean()) +
feature_extraction.TargetAgg(by='station', how=stats.EWMean(
0.5)) # aggregate feature (station and time)
)
model |= preprocessing.StandardScaler()
model |= linear_model.LinearRegression()
evaluate.progressive_val_score(dataset=datasets.Bikes(),
model=model,
metric=metrics.MAE(),
moment='moment',
delay=dt.timedelta(minutes=30),
print_every=20_000)
for x, y, in itertools.islice(X_y, 10000):
import copy
import functools
import math
import random
import numpy as np
import pytest
from river import stats
@pytest.mark.parametrize(
"stat",
[
pytest.param(stat, id=stat.__class__.__name__) for stat in
[stats.Mean(), stats.Var(
ddof=0), stats.Var(ddof=1)]
],
)
def test_add(stat):
A = copy.deepcopy(stat)
B = copy.deepcopy(stat)
C = copy.deepcopy(stat)
X = [random.random() for _ in range(30)]
Y = [random.random() for _ in range(30)]
W = [random.random() for _ in range(30)]
for x, y, w in zip(X, Y, W):
A.update(x, w)
B.update(y, w)
assert isinstance(pickle.loads(pickle.dumps(stat)), stat.__class__)
assert isinstance(copy.deepcopy(stat), stat.__class__)
# Check the statistic has a working __str__ and name method
assert isinstance(str(stat), str)
if isinstance(stat, stats.Univariate):
assert isinstance(stat.name, str)
@pytest.mark.parametrize(
'stat, func',
[(stats.Kurtosis(bias=True), sp_stats.kurtosis),
(stats.Kurtosis(bias=False),
functools.partial(sp_stats.kurtosis, bias=False)),
(stats.Mean(), statistics.mean), (stats.Skew(bias=True), sp_stats.skew),
(stats.Skew(bias=False), functools.partial(sp_stats.skew, bias=False)),
(stats.Var(ddof=0), np.var),
(stats.Var(), functools.partial(np.var, ddof=1))])
def test_univariate(stat, func):
# Shut up
np.warnings.filterwarnings('ignore')
X = [random.random() for _ in range(30)]
for i, x in enumerate(X):
stat.update(x)
if i >= 1:
assert math.isclose(stat.get(), func(X[:i + 1]), abs_tol=1e-10)
if isinstance(stat, stats.Univariate):
assert isinstance(stat.name, str)
@pytest.mark.parametrize("stat", load_stats(), ids=lambda stat: stat.__class__.__name__)
def test_repr_with_no_updates(stat):
assert isinstance(repr(stat), str)
assert isinstance(str(stat), str)
@pytest.mark.parametrize(
"stat, func",
[
(stats.Kurtosis(bias=True), sp_stats.kurtosis),
(stats.Kurtosis(bias=False), functools.partial(sp_stats.kurtosis, bias=False)),
(stats.Mean(), statistics.mean),
(stats.Skew(bias=True), sp_stats.skew),
(stats.Skew(bias=False), functools.partial(sp_stats.skew, bias=False)),
(stats.Var(ddof=0), np.var),
(stats.Var(), functools.partial(np.var, ddof=1)),
],
)
def test_univariate(stat, func):
# Shut up
np.warnings.filterwarnings("ignore")
X = [random.random() for _ in range(30)]
for i, x in enumerate(X):
stat.update(x)
def _unit_test_params(cls):
yield {"statistic": stats.Mean()}
def __init__(self, seed=None):
super().__init__()
self.variance = stats.Var()
self.mean = stats.Mean()
self.seed = seed
self._rng = random.Random(seed)
def __init__(self):
self.mean = stats.Mean()
def detail_eval(self, model, dataset, threshold=1.5):
"""
Divide input dataset relations into different categories (i.e. ONE-TO-ONE, ONE-TO-MANY,
MANY-TO-ONE and MANY-TO-MANY) according to the mapping properties of relationships.
Reference:
1. [Bordes, Antoine, et al. "Translating embeddings for modeling multi-relational data." Advances in neural information processing systems. 2013.](http://papers.nips.cc/paper/5071-translating-embeddings-for-modeling-multi-relational-data.pdf)
"""
mapping_type_relations = self.types_relations(model=model,
dataset=dataset,
threshold=threshold)
mapping_type_relations = {
self.relations[key]: value
for key, value in mapping_type_relations.items()
}
types_relations = ["1_1", "1_M", "M_1", "M_M"]
metrics = collections.OrderedDict({
"head-batch":
collections.OrderedDict({}),
"tail-batch":
collections.OrderedDict({})
})
for mode in ["head-batch", "tail-batch"]:
for type_relation in types_relations:
metrics[mode][type_relation] = collections.OrderedDict({
f"{metric}": stats.Mean()
for metric in ["MRR", "MR", "HITS@1", "HITS@3", "HITS@10"]
})
with torch.no_grad():
for test_set in self.get_entity_stream(dataset):
metrics = self.compute_detailled_score(
model=model,
test_set=test_set,
metrics=metrics,
types_relations=mapping_type_relations,
device=self.device,
)
for mode in ["head-batch", "tail-batch"]:
for type_relation in types_relations:
for metric in ["MRR", "MR", "HITS@1", "HITS@3", "HITS@10"]:
metrics[mode][type_relation][metric] = round(
metrics[mode][type_relation][metric].get(), 4)
results = pd.DataFrame(metrics)
head = pd.DataFrame(results["head-batch"].values.tolist())
tail = pd.DataFrame(results["tail-batch"].values.tolist())
head.columns = pd.MultiIndex.from_product([["head"], head.columns])
tail.columns = pd.MultiIndex.from_product([["tail"], tail.columns])
results = pd.concat([head, tail], axis="columns")
results = results.set_index(pd.Series(["1_1", "1_M", "M_1", "M_M"]))
results.index.name = "relation"
# Add frequency of each type of relation:
frequency = collections.OrderedDict()
for type_relation in types_relations:
frequency[type_relation] = 0
for _, type_relation in mapping_type_relations.items():
frequency[type_relation] += 1
for type_relation in types_relations:
frequency[type_relation] /= len(mapping_type_relations)
frequency = pd.DataFrame.from_dict(frequency,
orient="index",
columns=["frequency"])
frequency.columns = pd.MultiIndex.from_product([["metadata"],
frequency.columns])
results = pd.concat([results, frequency], axis="columns")
return results