电子游戏销售之回归模型与数据可视化

video_games_data = video_games.iloc[:, 5:8]
video_games_target = video_games.iloc[:, 9:10]
print(video_games_data)
print(video_games_target)

video_games_train, video_games_test, video_games_target, video_games_target_test = train_test_split(
    video_games_data, video_games_target)
clf = LinearRegression().fit(video_games_train, video_games_target) # 线性回归模型
print(clf)

video_games_target_test_pred=clf.predict(video_games_test)
print(video_games_target_test_pred[:20])
rcParams['font.sans-serif'] = 'SimHei'
fig = plt.figure(figsize=(10,6)) ##设定空白画布，并制定大小
##用不同的颜色表示不同数据
plt.plot(range(video_games_test.shape[0]),video_games_target_test,color="blue", linewidth=1.5, linestyle="-")
plt.plot(range(video_games_test.shape[0]),video_games_target_test_pred,color="red", linewidth=1.5, linestyle="-.")
plt.legend(['真实值','预测值'])

print('video_games数据线性回归模型1的平均误差为', mean_absolute_error(video_games_target_test1,video_games_target_test_pred1))
print('video_games数据线性回归模型1的均方误差', mean_squared_error(video_games_target_test1,video_games_target_test_pred1))
print('video_games数据线性回归模型1的中值绝对误差为', median_absolute_error(video_games_target_test1,video_games_target_test_pred1))
print('video_games数据线性回归模型1的可解释方差值为 ', explained_variance_score(video_games_target_test1,video_games_target_test_pred1))
print('video_games数据线性回归模型1的R²值', r2_score(video_games_target_test1,video_games_target_test_pred1))

highest_number_of_sales = df.groupby('Genre').sum(numeric_only=True).astype('int')
plt.figure(figsize=(24, 14)) # figuring the size
# makes bar plot
sns.barplot( # barplot
    x=highest_number_of_sales.index, # x-axis
    y=highest_number_of_sales["Global_Sales"], # y-axis
    data=highest_number_of_sales, # data
    palette="deep" # palette
)
plt.title( # title
    "不同类型的游戏销售量",
    weight="bold", # weight
    fontsize=35, # font-size
    pad=30 # padding
)
plt.xlabel( # x-label
    "类别",
    weight="bold", # weight
    color="purple", # color
    fontsize=25, # font-size
)
plt.xticks( # x-ticks
    weight="bold", # weight
    fontsize=15, # font-size
    rotation=10
)
plt.ylabel( # y-label
    "Global Sales",
    weight="bold", # weight
    color="green", # color
    fontsize=20 # font-size
)
plt.yticks( # y-ticks
    weight="bold", # weight
    fontsize=15 # font-size
)
plt.show()

corrmat= df.corr(numeric_only=True)
mask = np.zeros_like(corrmat)
mask[np.triu_indices_from(mask)] = True
with sns.axes_style("white"):
    f, ax = plt.subplots(figsize=(9,6))
    plt.title("相关性Correlation", weight="bold", fontsize=20, pad=30) # title
    plt.xticks(weight="bold", fontsize=10, rotation=45) # x-ticks
    plt.yticks(weight="bold", fontsize=10) # y-ticks
    plt.xlabel('各地区销售')
    plt.ylabel('发行年份')
    sns.heatmap(corrmat, cmap="YlGnBu", annot=True, mask=mask, square=True)
    plt.show()

rel = df.groupby(['Genre']).count().iloc[:,0]
rel = pd.DataFrame(rel.sort_values(ascending=False))
genres = rel.index
rel.columns = ['Releases']
colors = sns.color_palette("summer", len(rel))
plt.figure(figsize=(12,8))
ax = sns.barplot(y = genres , x = 'Releases', data=rel,orient='h', palette=colors)
ax.set_xlabel(xlabel='发行数量', fontsize=16)
ax.set_ylabel(ylabel='Genre', fontsize=16)
ax.set_title(label='按发行的游戏总数划分的类型', fontsize=20)
ax.set_yticklabels(labels = genres, fontsize=14)
plt.show()