Questions tagged «least-absolute-deviations»

2
如何用单纯形法求解最小绝对偏差?
argminwL(w)=∑ni=1|yi−wTx|arg⁡minwL(w)=∑i=1n|yi−wTx| \underset{\textbf{w}}{\arg\min} L(w)=\sum_{i=1}^{n}|y_{i}-\textbf{w}^T\textbf{x}| min∑ni=1uimin∑i=1nui\min \sum_{i=1}^{n}u_{i} ui≥xTw−yii=1,…,nui≥xTw−yii=1,…,nu_i \geq \textbf{x}^T\textbf{w}- y_{i} \; i = 1,\ldots,n ui≥−(xTw−yi)i=1,…,nui≥−(xTw−yi)i=1,…,nu_i \geq -\left(\textbf{x}^T\textbf{w}-y_{i}\right) \; i = 1,\ldots,n 但是我不知道要逐步解决它,因为我是LP的新手。你有什么主意吗?提前致谢! 编辑: 这是我已解决此问题的最新阶段。我正在尝试按照以下说明解决问题: 步骤1:将其制成标准格式 minZ=∑ni=1uiminZ=∑i=1nui\min Z=\sum_{i=1}^{n}u_{i} xTw−ui+s1=yii=1,…,nxTw−ui+s1=yii=1,…,n \textbf{x}^T\textbf{w} -u_i+s_1=y_{i} \; i = 1,\ldots,n xTw+ui+s2=−yii=1,…,nxTw+ui+s2=−yii=1,…,n \textbf{x}^T\textbf{w} +u_i+s_2=-y_{i} \; i = 1,\ldots,n 服从s1≥0;s2≥0;ui≥0 i=1,...,ns1≥0;s2≥0;ui≥0 i=1,...,ns_1 \ge 0; s_2\ge 0; u_i \ge 0 …
By using our site, you acknowledge that you have read and understand our Cookie Policy and Privacy Policy.
Licensed under cc by-sa 3.0 with attribution required.