def t_value_calc():
# 変数の確認用
def var_check(v):
print(v)
# 準備
np.random.seed(1)
resb = np.zeros(10000)
# 正規分布のインスタンス
norm_dist = sts.norm(loc=4, scale=0.8)
# シミュレーション
for i in range(0, 10000):
sample = norm_dist.rvs(size=10)
sample_mean = sp.mean(sample)
sample_std = sp.std(sample, ddof=1)
sample_se = sample_std / sp.sqrt(len(sample))
resb[i] = (sample_mean - 4) / sample_se
#
# var_check(resb)
#
title = "t_value_calc_graph and probability_density_graph"
plt.title(title)
# t値のヒストグラム
graph = sns.distplot(resb, color="black")
# 標準正規分布の確率密度
x = np.arange(start=-8, stop=8.1, step=0.1)
graph = plt.plot(x, sts.norm.pdf(x=x), color="black", linestyle="dotted")
canvas = f.image_graph(dt_name=title, dt_graph=graph)
canvas.view_option(me=4, st=0.8, va=0.64)
def arange_samplesize():
# シード値の固定
np.random.seed(1)
# samplesize=10
size_10 = f.sigma_calc(size=10, n_trial=10000)
size_10_df = pd.DataFrame({
"list_mean": size_10.create_source(),
"size": np.tile("size 10", 10000)
})
# samplesize=20
size_20 = f.sigma_calc(size=20, n_trial=10000)
size_20_df = pd.DataFrame({
"list_mean": size_20.create_source(),
"size": np.tile("size 20", 10000)
})
# samplesize=30
size_30 = f.sigma_calc(size=30, n_trial=10000)
size_30_df = pd.DataFrame({
"list_mean": size_30.create_source(),
"size": np.tile("size 30", 10000)
})
# 結合
sim_result = pd.concat([size_10_df, size_20_df, size_30_df])
# print(sim_result.head())
# グラフの表示
title = "arange_samplesize_graph"
plt.title(title)
graph = sns.violinplot(x="size",
y="list_mean",
data=sim_result,
color="gray")
canvas = f.image_graph(graph, title)
canvas.view_option(me="", st="", va="")
def graph_plot(gh_title, gh_source):
# tmp_me = input("平均値(特になければ空でOK) >>> ")
# tmp_st = input("標準偏差(特になければ空でOK) >>> ")
# tmp_va = input("分散(特になければ空でOK) >>> ")
plt.title(gh_title)
canvas = f.image_graph(dt_name=gh_title, dt_graph=gh_source)
canvas.view_option(me="", st="", va="")
def sigma_graph(list, op1, op2, op3):
title = "fish_population_graph"
plt.title(title)
graph = sns.distplot(list, kde=False, color='black')
# 表示
canvas = f.image_graph(graph, title)
canvas.view_option(op1, op2, op3)
def sigma_standard_error():
# prepare
spm = np.arange(start=2, stop=102, step=2)
std_box = np.zeros(len(spm))
# 単純な計算式
st_err_1 = 0.8 / np.sqrt(spm)
# 標本平均の標準偏差
np.random.seed(1)
for i in range(0, len(spm)):
tmp = f.sigma_calc(size=spm[i], n_trial=100)
std_box[i] = sp.std(tmp.create_source(), ddof=1)
# 変数の確認用
def var_conf(v):
print(v)
# var_conf(std_box) # コメントアウト推奨
# 標本平均の標準偏差と標準誤差のグラフ
title = "sigma_standard_error_graph"
plt.title(title)
graph = plt.plot(spm, std_box, color="black")
graph = plt.plot(spm, st_err_1, linestyle="dotted", color="black")
graph = plt.xlabel("smaple_size")
graph = plt.ylabel("mean_std_value")
canvas = f.image_graph(dt_graph=graph, dt_name=title)
canvas.view_option(me="", st="", va="")
def tbunpu_and_tvalue():
#--- t値（標本から計算された標準誤差）
np.random.seed(1)
resb = np.zeros(10000)
# 正規分布のインスタンス
norm_dist = sts.norm(loc=4, scale=0.8)
# シミュレーション
for i in range(0, 10000):
sample = norm_dist.rvs(size=10)
sample_mean = sp.mean(sample)
sample_std = sp.std(sample, ddof=1)
sample_se = sample_std / sp.sqrt(len(sample))
resb[i] = (sample_mean - 4) / sample_se
#--- まで ---
#--- t分布の確率密度
x = np.arange(start=-8, stop=8.1, step=0.1)
#--- まで
# グラフの描写
title = "tbunpu_and_tvalue_graph"
plt.title(title)
sns.distplot(resb, color="black", norm_hist=True)
graph = plt.plot(x, sts.t.pdf(x=x, df=9), color="black", linestyle="dotted")
canvas = f.image_graph(dt_name=title, dt_graph=graph)
canvas.view_option(me="", st="", va="")
def like_graph(list, density, op1, op2, op3):
title = "fixed_population_graph"
plt.title(title)
graph = plt.plot(list, density, color="black")
# 表示
canvas = f.image_graph(graph, title)
canvas.view_option(op1, op2, op3)
def barplot(data):
# グラフデザインの指定
sns.set()
#グラフの描写
title = "barplot_graph"
plt.title(title)
graph = sns.barplot(x="species", y="length", data=data, color="gray")
#表示
canvas_view = f.image_graph(graph, title)
canvas_view.view()
def jointplot(data):
# グラフデザインの指定
sns.set()
#グラフの描写
title = "jointplot_graph"
plt.title(title)
graph = sns.jointplot(x="x", y="y", data=data, color="black")
#表示
canvas_view = f.image_graph(graph, title)
canvas_view.view()
def t_bunpu():
# 準備
x = np.arange(start=-8, stop=8.1, step=0.1)
# グラフの描写
title="t_bunpu_graph"
plt.title(title)
plt.plot(x, sts.norm.pdf(x=x), color="black", linestyle="dotted")
graph = plt.plot(x, sts.t.pdf(x=x, df=9), color="black")
canvas = f.image_graph(dt_name=title, dt_graph=graph)
canvas.view_option(va="", st="", me="")
def iris_graph(data):
#グラフデザインの指定
sns.set()
#グラフの描写
title = "iris_pairplot_graph"
plt.title(title)
graph = sns.pairplot(data, hue="species",
palette="gray") # hueはカテゴリ型データの列名を指定する
#表示
canvas_view = f.image_graph(graph, title)
canvas_view.view()
def kernel_design():
# データの用意
list_data = np.array([2, 3, 3, 4, 4, 4, 4, 5, 5, 6])
# グラフデザインの設定
sns.set()
# グラフの描写
title = "kernel_design_histgram"
plt.title(title)
graph = sns.distplot(list_data, color="black")
canvas_view = f.image_graph(graph, title)
canvas_view.view()
def bins_1_design():
# データの用意
list_data = np.array([2, 3, 3, 4, 4, 4, 4, 5, 5, 6])
# グラフデザインの設定
sns.set()
# グラフの描写
title = "bins_1_histgram"
plt.title(title)
graph = sns.distplot(list_data, bins=1, color="black", kde=False)
canvas_view = f.image_graph(graph, title)
canvas_view.view()
def histogram():
# データの用意
list_data = np.array([2, 3, 3, 4, 4, 4, 4, 5, 5, 6])
# グラフデザインの設定
sns.set()
# グラフの描写
title = "histogram_seaborn"
plt.title(title)
graph = sns.distplot(list_data, bins=5, color="black", kde=False)
canvas_view = f.image_graph(graph, title)
canvas_view.view()
def lineplot():
# データの用意
x = np.array([0, 1, 2, 3, 4, 5, 6, 7, 8, 9])
y = np.array([2, 3, 4, 3, 5, 4, 6, 7, 4, 8])
# グラフの描写
title = "linegraph"
plt.title(title)
graph = plt.plot(x, y, color="blue")
canvas_view = f.image_graph(graph, title)
canvas_view.view()
def lineplot_seaborn():
# データの用意
x = np.array([0, 1, 2, 3, 4, 5, 6, 7, 8, 9])
y = np.array([2, 3, 4, 3, 5, 4, 6, 7, 4, 8])
# グラフデザインの設定
sns.set()
# グラフの描写
title = "linegraph_seaborn"
plt.title(title)
graph = plt.plot(x, y, color="black")
canvas_view = f.image_graph(graph, title)
canvas_view.view()
def fish_fixed_graph():
# 準備
x = pd.read_csv("/root/app/sts4_csv.csv")["length"]
y = np.arange(start=1, stop=7.1, step=0.1)
# 表示
title = "fish_fixed_graph"
plt.title(title)
graph = sns.distplot(x, kde=False, norm_hist=True, color="black")
graph = plt.plot(y,
stats.norm.pdf(x=y, loc=4, scale=0.8),
color="black")
canvas = f.image_graph(graph, title)
canvas.view_option(4, 0.8, 0.64)
def multi_line(list_data):
# グループ別に分割
list_a = list_data.query('species=="A"')["length"]
list_b = list_data.query('species=="B"')["length"]
# グラフデザインの指定
sns.set()
#グラフの描写
title = "multi_line_graph"
plt.title(title)
graph = sns.distplot(list_a, bins=5, color="red", kde=False)
graph = sns.distplot(list_b, bins=5, color="blue", kde=False)
# グラフの表示
canvas_view = f.image_graph(graph, title)
canvas_view.view()
def coin_check():
n_size = 10000
n_trial = 50000
# 表なら1, 裏なら0
coin = np.array([0, 1])
count_coin = np.zeros(n_trial)
# 思考実験：コインをn_size回数投げる施工をn_trial回行う
np.random.seed(1)
for i in range(0, n_trial):
count_coin[i] = sp.sum(
sp.random.choice(coin, size=n_size, replace=True))
# ヒストグラムで描写
title = "coin_check_graph(ver.Histgram)"
plt.title(title)
graph = sns.distplot(count_coin, color="black")
canvas = f.image_graph(dt_name=title, dt_graph=graph)
canvas.view_option(me="", va="", st="")
def sigma_mean_feature(pop):
box = np.zeros(10000) # 空のリストを用意
np.random.seed(1) # 乱数の固定
# 標本抽出のシミュレーション
for i in range(0, 10000):
var = pop.rvs(size=10)
box[i] = sp.mean(var)
box_mean = np.mean(box) # １００００もの標本平均の平均
box_std = sp.std(box, ddof=1) # １００００もの標本平均の不偏標準偏差
box_var = sp.var(box, ddof=1) # ついでに分散も
# グラフに示す
title = "sigma_mean_feature_graph"
plt.title(title)
graph = sns.distplot(box, color="black")
canvas = f.image_graph(graph, title)
canvas.view_option(box_mean, box_std, box_var)
def sigma_mean_deviation():
# prepare
base_samplesize = np.arange(start=2, stop=102, step=2)
base_box = np.zeros(len(base_samplesize))
# シード値の固定
np.random.seed(1)
for i in range(0, len(base_samplesize)):
tmp_box = f.sigma_calc(size=base_samplesize[i], n_trial=100)
base_box[i] = sp.std(tmp_box.create_source(), ddof=1)
# グラフの描写
title = "sigma_mean_deviation_graph"
plt.title(title)
graph = plt.plot(base_samplesize, base_box, color="black")
graph = plt.xlabel("base_samplesize")
graph = plt.ylabel("base_box")
canvas = f.image_graph(graph, title)
canvas.view_option(me="", st="", va="")
def arange_samplesize_var():
# シード値の固定
np.random.seed(1)
# 準備
pop = stats.norm(loc=4, scale=0.8)
# 各サンプルサイズの用意
size = np.arange(start=10, stop=100010, step=10)
box = np.zeros(len(size))
for i in range(0, len(size)):
tmp = pop.rvs(size=size[i])
box[i] = sp.var(tmp, ddof=1)
# グラフの描写
title = "range_samplesize_var_graph"
plt.title(title)
plt.xlabel("sample_size")
plt.ylabel("arange_var_data")
graph = plt.plot(size, box, color="black")
canvas = f.image_graph(dt_graph=graph, dt_name=title)
canvas.view_option(me=4, st=0.8, va="")
def bigger_samplesize(pop):
# サンプルサイズの用意(10 ~ 100100まで100区切りでプールを用意)
size_array = np.arange(start=10, stop=100100, step=100)
# 標本平均を格納する箱の用意
samplesize_box = np.zeros(len(size_array))
# 標本平均をサンプルサイズ別に格納
np.random.seed(1)
for i in range(0, len(size_array)):
dt = pop.rvs(size=size_array[i])
samplesize_box[i] = sp.mean(dt)
# 標本平均の各情報
box_mean = sp.mean(samplesize_box)
box_std = sp.std(samplesize_box, ddof=1)
box_var = sp.var(samplesize_box, ddof=1)
# グラの描写
title = "bigger_samplesize_graph"
plt.title(title)
graph = plt.plot(size_array, samplesize_box, color="black")
graph = plt.xlabel("sanple_size")
graph = plt.ylabel("sample_mean")
canvas = f.image_graph(graph, title)
canvas.view_option(box_mean, box_std, box_var)
def debug_space():
test = f.image_graph("graph", "graph_title")
test.view_option("", "", "")
