課后答案上機(jī)練習(xí)答案：#導(dǎo)入必要庫importpandasaspdimportnumpyasnpimportmatplotlib.pyplotaspltfromsklearn.preprocessingimportStandardScalerfromsklearn.model_selectionimporttrain_test_split,learning_curvefromsklearn.ensembleimportRandomForestClassifierfromsklearn.metricsimportclassification_report,accuracy_scoreimportjoblib第1步，本地?cái)?shù)據(jù)集加載file_path='中醫(yī)證素.xls'TCM_Constitution_data=pd.read_excel(file_path)TCM_Constitution_data.head()第2步，描述性統(tǒng)計(jì)分析print("數(shù)據(jù)集的行數(shù)與列數(shù):{}".format(TCM_Constitution_data.shape))#數(shù)據(jù)形狀#繪制'y1'到'y2'列的柱狀圖#設(shè)置中文字體plt.rcParams['font.sans-serif']=['SimHei']#用來正常顯示中文標(biāo)簽plt.rcParams['axes.unicode_minus']=False#用來正常顯示負(fù)號(hào)#定義柱狀圖的顏色調(diào)色板colors=['blue','green','red','purple','orange','brown','pink','gray']#使用定義的顏色創(chuàng)建中醫(yī)體質(zhì)樣本數(shù)量統(tǒng)計(jì)柱狀圖TCM_Constitution_data.iloc[:,40:].sum().plot(kind='bar',color=colors,figsize=(5,3))plt.title('中醫(yī)體質(zhì)樣本數(shù)量統(tǒng)計(jì)')plt.xlabel('中醫(yī)體質(zhì)')plt.ylabel('樣本數(shù)量')plt.xticks(rotation=0)#將x軸標(biāo)簽旋轉(zhuǎn)為水平，以便于閱讀plt.tight_layout()#調(diào)整繪圖，確保所有內(nèi)容沒有重疊plt.show()第3步，數(shù)據(jù)處理（異常值處理、數(shù)據(jù)分割）#異常值處理#定義一個(gè)函數(shù)，使用IQR方法檢測(cè)并替換異常值為中位數(shù)defreplace_outliers_with_median(series):#計(jì)算四分位數(shù)Q1和Q3Q1=series.quantile(0.25)#下四分位數(shù)：數(shù)據(jù)集中25%的數(shù)據(jù)小于這個(gè)值Q3=series.quantile(0.75)#上四分位數(shù)：數(shù)據(jù)集中75%的數(shù)據(jù)小于這個(gè)值IQR=Q3-Q1#四分位數(shù)范圍（IQR）：上四分位數(shù)和下四分位數(shù)之間的距離#計(jì)算異常值的下界和上界lower_bound=Q1-1.5*IQR#異常值下界：Q1減去1.5倍的IQRupper_bound=Q3+1.5*IQR#異常值上界：Q3加上1.5倍的IQR#任何低于下界或高于上界的值都被視為異常值#計(jì)算中位數(shù)median=series.median()#中位數(shù)：數(shù)據(jù)集中的中間值#將超出上下界的異常值替換為中位數(shù)series[~series.between(lower_bound,upper_bound)]=median#這里使用了邏輯運(yùn)算符~來取反，.between(lower_bound,upper_bound)返回一個(gè)布爾型數(shù)組#表示每個(gè)值是否在上下界之間，取反后就是異常值的位置returnseries#應(yīng)用函數(shù)到x1到x39列x_columns=TCM_Constitution_data.filter(regex='x[0-9]+')#使用正則表達(dá)式篩選出x1到x39的列TCM_Constitution_data[x_columns.columns]=x_columns.apply(replace_outliers_with_median)#對(duì)篩選出的列應(yīng)用替換異常值的函數(shù)#顯示處理異常值后的前幾行數(shù)據(jù)TCM_Constitution_data.head()#分割數(shù)據(jù)集X=TCM_Constitution_data[['x1','x2','x3','x4','x5','x6','x7','x8','x9','x10','x11','x12','x13','x14','x15','x16','x17','x18','x19','x20','x21','x22','x23','x24','x25','x26','x27','x28','x29','x30','x31','x32','x33','x34','x35','x36','x37','x38','x39']]y=TCM_Constitution_data[['y1','y2']]X_train,X_test,y_train,y_test=train_test_split(X,y,test_size=0.2,random_state=42)X_train.shape,X_test.shape,y_train.shape,y_test.shape第4步，模型訓(xùn)練model=RandomForestClassifier(random_state=42)model.fit(X_train,y_train)第5步，模型評(píng)估（模型測(cè)試、繪制學(xué)習(xí)曲線、繪制訓(xùn)練驗(yàn)證得分曲線、隨機(jī)抽取樣例驗(yàn)證）y_pred=model.predict(X_test)print(classification_report(y_test,y_pred))print(f'準(zhǔn)確率:{accuracy_score(y_test,y_pred):.3f}')#使用learning_curve函數(shù)獲取訓(xùn)練和測(cè)試分?jǐn)?shù)train_sizes,train_scores,test_scores=learning_curve(model,X,y,cv=5,n_jobs=-1,train_sizes=np.linspace(0.1,1.0,10))#計(jì)算訓(xùn)練和驗(yàn)證分?jǐn)?shù)的平均值和標(biāo)準(zhǔn)差train_scores_mean=np.mean(train_scores,axis=1)train_scores_std=np.std(train_scores,axis=1)test_scores_mean=np.mean(test_scores,axis=1)test_scores_std=np.std(test_scores,axis=1)#繪制學(xué)習(xí)曲線plt.figure()plt.title("學(xué)習(xí)曲線")plt.xlabel("訓(xùn)練樣本數(shù)量")plt.ylabel("分?jǐn)?shù)")plt.grid()#填充訓(xùn)練和驗(yàn)證分?jǐn)?shù)的平均值與標(biāo)準(zhǔn)差的區(qū)域plt.fill_between(train_sizes,train_scores_mean-train_scores_std,train_scores_mean+train_scores_std,alpha=0.1,color="r")plt.fill_between(train_sizes,test_scores_mean-test_scores_std,test_scores_mean+test_scores_std,alpha=0.1,color="g")#繪制訓(xùn)練和驗(yàn)證分?jǐn)?shù)的曲線plt.plot(train_sizes,train_scores_mean,'o-',color="r",label="訓(xùn)練分?jǐn)?shù)")plt.plot(train_sizes,test_scores_mean,'o-',color="g",label="交叉驗(yàn)證分?jǐn)?shù)")#添加圖例plt.legend(loc="best")#顯示圖表plt.show()#隨機(jī)選取五個(gè)樣例查看原始結(jié)果和預(yù)測(cè)結(jié)果#設(shè)置隨機(jī)樣本的數(shù)量num_samples=5#生成隨機(jī)索引random_indices=np.random.randint(0,len(X_test),num_samples)#遍歷隨機(jī)索引，獲取每個(gè)樣本及其原始和預(yù)測(cè)結(jié)果foriinrandom_indices:#獲取測(cè)試集中的樣本sample=X_test.iloc[i]#獲取原始的測(cè)試集標(biāo)簽original_outcome=y_test.iloc[i]#使用模型進(jìn)行預(yù)測(cè)predi