多元線性回歸—C語言

1、多元線性回歸算C語言上篇已經(jīng)介紹過簡單線性回歸了，這篇介紹第二個算法，多元線性回歸，僅介紹兩個主要函數(shù)：預(yù)測函數(shù)和訓(xùn)練函數(shù)。算法介紹模型可以如下表示：y=bo+b1xx1+b2xx2+.隨機(jī)梯度下降法多元線性回歸采用的是隨機(jī)梯度下降方法，該方法具體介紹同樣自行百度，在這里不做詳細(xì)介紹。僅介紹C語言實(shí)現(xiàn)方法這里給出更新方程：b=b-learningratexerrorxx函數(shù)讀取“以下三個函數(shù)分別為獲取行數(shù)、獲取列數(shù)、獲取文本內(nèi)容。double*dataset;introw,col;intget_row(char*filename)獲取行數(shù)charline1024;inti=0;FILE*str

2、eam=fopen(filename,r);while(fgets(line,1024,stream)i+;fclose(stream);returni;獲取列數(shù)charline1024;inti=0;FILE*stream=fopen(filename,r);fgets(line,1024,stream);char*token=strtok(line,);while(token)token=strtok(NULL,);i+;fclose(stream);returni;voidget_two_dimension(char*line,double*data,char*filename)FILE

3、*stream=fopen(filename,r);inti=0;while(fgets(line,1024,stream)/逐行讀取intj=0;char*tok;char*tmp=strdup(line);for(tok=strtok(line,);tok&*tok;j+,tok=strtok(NULL,n)dataij=atof(tok);/轉(zhuǎn)換成浮點(diǎn)數(shù)/字符串拆分操作i+;free(tmp);fclose(stream);/文件打開后要進(jìn)行關(guān)閉操作EXAMPLEintmain()charfilename=data.csv;charline1024;double*data;introw,

4、col;row=get_row(filename);col=get_col(filename);data=(double*)malloc(row*sizeof(int*);for(inti=0;irow;+i)datai=(double*)malloc(col*sizeof(double);動態(tài)申請二維數(shù)組get_two_dimension(line,data,filename);printf(row=%dn,row);printf(col=%dn,col);inti,j;for(i=0;irow;i+)for(j=O;jvcol;j+)floatmean=(float)(sum/length

5、);returnmean;隨機(jī)梯度下降估計(jì)回歸系數(shù)更新方程：KaTeXparseerror:Undefinedcontrolsequenee:atposition29:iction-expected、b1(t+1)=b1(t)一learningratexerror(t)xx1(t)bo(t+1)=bo(t)一learningratexerror(t)/參數(shù)為：數(shù)據(jù)集、每個數(shù)據(jù)集屬性個|帶label)、存放系數(shù)的數(shù)組、學(xué)習(xí)率、epoch、train_sizedouble*coefficients_sgd(double*dataset,intcol,doublecoef,doublel_rate,

6、intn_epoch,inttrain_size)inti;for(i=0;in_epoch;i+)intj=0;/遍歷每一行for(j=0;jtrain_size;j+)doubleyhat=predict(col,datasetj,coef);doubleerr=yhat-datasetjcol-1;coef0-=l_rate*err;intk;for(k=0;kcol-1;k+)coefk+1-=l_rate*err*datasetjk;for(i=0;icol;i+)printf(coefi=%fn,coefi);returncoef;預(yù)測函數(shù)參數(shù)：樣本屬性個數(shù)、樣本、系數(shù)doublepredict(intcol,doublearray,doublecoefficien