 GitHub でソースを表示 |
1 年後のクレジット カードの金利を見積もりたいと想像してみましょう。現在のプライム レートが 2% で、クレジット カード会社が 10% にプライムを加えた料金を請求するとします。現在の経済の好調さを考慮すると、あなたは連邦準備制度が利上げをしないよりも利上げをする可能性の方が高いと考えています。 FRBは今後12カ月間に8回会合を開き、フェデラルファンド金利を0.25％引き上げるか、以前の水準に据え置くかのどちらかとなる。
二項分布を使用して、12 か月期間終了時のクレジット カードの金利をモデル化します。具体的には、次のパラメータを持つ TensorFlow 確率二項分布クラスを使用します: total_count = 8 (試行または会議の数)、probs = {.6, .7, .8, .9}、についての推定範囲についてFRBが各会合でフェデラルファンド金利を0.25%引き上げる確率。
依存関係と前提条件
TensorFlow 確率インストール設定
TFP_Installation = "Stable TFP"
if TFP_Installation == "Most Recent TFP":
!pip install -q tfp-nightly
print("Most recent TFP version installed")
elif TFP_Installation == "Stable TFP":
!pip install -q --upgrade tensorflow-probability
print("Up-to-date, stable TFP version installed")
elif TFP_Installation == "Stable TFP-GPU":
!pip install -q --upgrade tensorflow-probability-gpu
print("Up-to-date, stable TFP-GPU version installed")
print("(make sure GPU is properly configured)")
elif TFP_Installation == "Most Recent TFP-GPU":
!pip install -q tfp-nightly-gpu
print("Most recent TFP-GPU version installed")
print("(make sure GPU is properly configured)")
elif TFP_Installation == "TFP Already Installed":
print("TFP already installed in this environment")
pass
else:
print("Installation Error: Please select a viable TFP installation option.")
インポートとグローバル変数 (必ずこのセルを実行してください)
from __future__ import absolute_import, division, print_function
warning_status = "ignore"
import warnings
warnings.filterwarnings(warning_status)
with warnings.catch_warnings():
warnings.filterwarnings(warning_status, category=DeprecationWarning)
warnings.filterwarnings(warning_status, category=UserWarning)
import numpy as np
import os
matplotlib_style = 'fivethirtyeight'
import matplotlib.pyplot as plt; plt.style.use(matplotlib_style)
import matplotlib.axes as axes;
from matplotlib.patches import Ellipse
%matplotlib inline
import seaborn as sns; sns.set_context('notebook')
notebook_screen_res = 'png'
%config InlineBackend.figure_format = notebook_screen_res
import tensorflow as tf
# Eager Execution
use_tf_eager = True
# Use try/except so we can easily re-execute the whole notebook.
if use_tf_eager:
try:
tf.compat.v1.enable_eager_execution()
except:
reset_session()
import tensorflow_probability as tfp
tfd = tfp.distributions
tfb = tfp.bijectors
def default_session_options(enable_gpu_ram_resizing=True,
enable_xla=False):
"""Creates default options for Graph-mode session."""
config = tf.ConfigProto()
config.log_device_placement = True
if enable_gpu_ram_resizing:
# `allow_growth=True` makes it possible to connect multiple
# colabs to your GPU. Otherwise the colab malloc's all GPU ram.
config.gpu_options.allow_growth = True
if enable_xla:
# Enable on XLA. https://www.tensorflow.org/performance/xla/.
config.graph_options.optimizer_options.global_jit_level = (
tf.OptimizerOptions.ON_1)
return config
def reset_session(options=None):
"""Creates a new global, interactive session in Graph-mode."""
if tf.executing_eagerly():
return
global sess
try:
tf.reset_default_graph()
sess.close()
except:
pass
if options is None:
options = default_session_options()
sess = tf.InteractiveSession(config=options)
def evaluate(tensors):
"""Evaluates Tensor or EagerTensor to Numpy `ndarray`s.
Args:
tensors: Object of `Tensor` or EagerTensor`s; can be `list`, `tuple`,
`namedtuple` or combinations thereof.
Returns:
ndarrays: Object with same structure as `tensors` except with `Tensor` or
`EagerTensor`s replaced by Numpy `ndarray`s.
"""
if tf.executing_eagerly():
return tf.contrib.framework.nest.pack_sequence_as(
tensors,
[t.numpy() if tf.contrib.framework.is_tensor(t) else t
for t in tf.contrib.framework.nest.flatten(tensors)])
return sess.run(tensors)
class _TFColor(object):
"""Enum of colors used in TF docs."""
red = '#F15854'
blue = '#5DA5DA'
orange = '#FAA43A'
green = '#60BD68'
pink = '#F17CB0'
brown = '#B2912F'
purple = '#B276B2'
yellow = '#DECF3F'
gray = '#4D4D4D'
def __getitem__(self, i):
return [
self.red,
self.orange,
self.green,
self.blue,
self.pink,
self.brown,
self.purple,
self.yellow,
self.gray,
][i % 9]
TFColor = _TFColor()
確率の計算
12 か月以内に発生する可能性のあるクレジット カードの金利の確率を計算します。
# First we encode our assumptions.
num_times_fed_meets_per_year = 8.
possible_fed_increases = tf.range(
start=0.,
limit=num_times_fed_meets_per_year + 1)
possible_cc_interest_rates = 2. + 10. + 0.25 * possible_fed_increases
prob_fed_raises_rates = tf.constant([0.6, 0.7, 0.8, 0.9]) # Wild guesses.
# Now we use TFP to compute probabilities in a vectorized manner.
# Pad a dim so we broadcast fed probs against CC interest rates.
prob_fed_raises_rates = prob_fed_raises_rates[..., tf.newaxis]
prob_cc_interest_rate = tfd.Binomial(
total_count=num_times_fed_meets_per_year,
probs=prob_fed_raises_rates).prob(possible_fed_increases)
TFコードを実行する
# Convert from TF to numpy.
[
possible_cc_interest_rates_,
prob_cc_interest_rate_,
prob_fed_raises_rates_,
] = evaluate([
possible_cc_interest_rates,
prob_cc_interest_rate,
prob_fed_raises_rates,
])
結果を視覚化する
plt.figure(figsize=(14, 9))
for i, pf in enumerate(prob_fed_raises_rates_):
plt.subplot(2, 2, i+1)
plt.bar(possible_cc_interest_rates_,
prob_cc_interest_rate_[i],
color=TFColor[i],
width=0.23,
label="$p = {:.1f}$".format(pf[0]),
alpha=0.6,
edgecolor=TFColor[i],
lw="3")
plt.xticks(possible_cc_interest_rates_ + 0.125, possible_cc_interest_rates_)
plt.xlim(12, 14.25)
plt.ylim(0, 0.5)
plt.ylabel("Probability of cc interest rate")
plt.xlabel("Credit card interest rate (%)")
plt.title("Credit card interest rates: "
"prob_fed_raises_rates = {:.1f}".format(pf[0]));
plt.suptitle("Estimates of credit card interest rates in 12 months.",
fontsize="x-large",
y=1.02)
plt.tight_layout()
