首页 > 移动平台 > 详细

线性代数《Linear Algebra and Its Application》学习总结

时间:2014-05-07 00:08:25      阅读:725      评论:0      收藏:0      [点我收藏+]

此文仅为学习记录,内容会包括一些数学概念,定义,个人理解的摘要。希望能够分享一些学习内容。

第一节:Row Reduction and Echelon Forms

  1. Echelon form: 行消元后的矩阵
  2. Reduced echelon form: 行消元并且leading entry为1的矩阵。
  3. Echelon form and reduced echelon form are row equivalent to the original form.
  4. Span{v1, v2, v3,...... vp} is the collection of all vectors that can be written in the form c1*v1 + c2*v2 + ...... cp*vp with c1, .... cp scalars.
  5. Ax = 0 has a nontrival solution if and only if the equation has at least one free variable.(not full column rank)
  6. Ax = b 的解等于 Ax = 0 和 特解的和。
  7. 解线性方程组流程P54。
  8. 线性无关指任何向量不能组合成其中一个向量。
  9. Ax = b : ColA1 * x1 + ColA2 * x2 +.... ColAm * xm = b
  10. Matrix Transformations: T(x) = Ax is linear transformation.
  11. 转换矩阵是各维单位转换后的组合。A = [T(e1) T(e2) .. T(en)]
  12. A mapping T: R^n -> R^m is said to be onto R^m if each b in R^m is the image of at least one x in R^n. (Ax = b 有解)
  13. A mapping T: R^n -> R^m is said to be one-to-one R^m if each b in R^m is the image of at most one x in R^n.

第二节:Matrix Operation

  1. Each column of AB is a linear combination of the columns of A using weightings from the corresponding columns of B. AB = A[b1  b2 b3 b4 ,,, bp] = [Ab1 Ab2 ... Abp]
  2. Each row of AB is a linear combination of the columns of B using weightings from the corresponding rows of A.
  3. Warning: AB != BA. AB = AC !=> B = C. AB = 0 !=> A = 0 or B = 0
  4. 逆矩阵的定义:A-1*A = A*A-1 = E. 可以推导出A为方阵,详见Exercise 23-25 ,Section 2.1. A可逆的充要条件为A满秩(行列式不等于0)。
  5. 对[A I] 做行消元可以得到[I A-1]
  6. 矩阵满秩的所有等价定义:P129,P179.
  7. LU分解:A = LU,其中L为对角元素为1,的下半方阵,U为m*n的上半矩阵。L为变换矩阵的乘机的逆,U为A的Echelon form。计算L不需要计算各变换矩阵。详见P146。
  8. subspace, column space, null space的定义。
  9. A = m*n => rank(A) + rank(Nul(A)) = n.
  10. The dimension of a nonzero subspace H, denoted by dim H, is the numbers of vectors in any basis for H. The dimension of the zero subspace {0} us defined to be zero.

第三节:Introduction to Determinants

  1. determinant的定义和计算方式。
  2. 行消元不改变行列式值。交换行改变正负号。某一行乘以k,那么行列式乘以k。
  3. 三角矩阵的行列式为对角元素的乘积。
  4. det(AB) = det(A) * det(B)。
  5. Let A be an invertible n*n matrix. For any b in R^n, the unique solutionx of Ax = b has entries given by xi = det Ai(b)/det(A)。 Ai(b) 表示用b替换A的第i行。
  6. 由5可以推导出A^-1 = 1/det(A) * adj A. adj A = [(-1)^i+j* det(Aji)]
  7. 行列式与体积的关系:平行几何体的面积或者体积等于|det(A)|。而且 det(Ap) = det(A)*det(p)

第四节:Vector Spaces

  1. An indexed set {v1, v2, ... ... vp} of two or more vectors, with vi != 0, is linearly dependent, if and only if some vj (with j > 1) is a linear combination of the preceding vectors.
  2. Elementary row operation  on a matrix do not affect the linear dependence relations among the columns of the matrix.
  3. Row operations can change the column space of a matrix.
  4. x = Pb [x]b: we call Pb the change-of-coordinates matrix from B to the standard basis in R^n.
  5. Let B and C be bases of a vector space V. Then there is a unique n*n matrix P_C<-B such that [x]c = P_C<-B [x]b. The columns of P_C<-B are the C-coordinate vectors of the vectors in the basis B, that is P_C<-B = [[b1]c [b2]c ... [bn]c]. [ C B ] ~ [ I P_C<-B]

第五节:Eigenvectors and Eigenvalues

  1. Ax=λ?xbubuko.com,布布扣
  2. 不同特征值对应的特征向量线性无关。
  3. det(A - λ *I) = 0. 因为(A - λ *I)有非零解。
  4. A is similar to B if there is an invertible matrix P such that P^-1AP = B. They have same eigenvalues.
  5. 矩阵能够对角化的条件是有n个线性无关的特征向量(特征向量有无穷多个,线性无关向量的数量最多为n)。
  6. 特征空间的维度小于等于特征根的幂。当特征空间的维度等于特征根的幂,矩阵能够对角化。
  7. 相同坐标变换矩阵在不同维度空间坐标系下的转换:P328。相同坐标变换矩阵在不同坐标系的转换:P329。其实都是一样的。
  8. Suppose A = PDP^-1, where D is a diagonal n*n matrix. If B is the basis for R^n formed from the columns of P, then D is the B-matrix for the transformation x ->Ax. 当坐标系转换为P时,转换矩阵对应变成对角矩阵。
  9. 复数系统。
  10. 迭代求特征值和特征向量。 先估计一个特近的特征值和一个向量xbubuko.com,布布扣0bubuko.com,布布扣bubuko.com,布布扣 (其中的最大元素为1)。然后迭代,迭代流程详见P365。迭代可以得到最大特征值的原因如下:因为(λbubuko.com,布布扣1bubuko.com,布布扣)bubuko.com,布布扣?kbubuko.com,布布扣Abubuko.com,布布扣kbubuko.com,布布扣xcbubuko.com,布布扣1bubuko.com,布布扣vbubuko.com,布布扣1bubuko.com,布布扣bubuko.com,布布扣 ,所以对于任意xbubuko.com,布布扣 ,当k趋近无穷的时候,Abubuko.com,布布扣kbubuko.com,布布扣xbubuko.com,布布扣 会和特征向量同向。虽然λbubuko.com,布布扣 cbubuko.com,布布扣1bubuko.com,布布扣vbubuko.com,布布扣1bubuko.com,布布扣bubuko.com,布布扣 都未知,但是由于Axbubuko.com,布布扣kbubuko.com,布布扣bubuko.com,布布扣 会趋近λ?xbubuko.com,布布扣kbubuko.com,布布扣bubuko.com,布布扣 ,我们只要令xbubuko.com,布布扣kbubuko.com,布布扣bubuko.com,布布扣 的最大元素为1,就能得到λbubuko.com,布布扣

 

第六节 :Inner Product, Length, and Orthogonality

  1. (RowA)bubuko.com,布布扣bubuko.com,布布扣=NulAbubuko.com,布布扣 and (ColA)bubuko.com,布布扣bubuko.com,布布扣=NulAbubuko.com,布布扣?bubuko.com,布布扣bubuko.com,布布扣 . 这很显然,其中Abubuko.com,布布扣bubuko.com,布布扣bubuko.com,布布扣 表示与A空间垂直的空间。
  2. An orthogonal basis for a subspace W of Rbubuko.com,布布扣nbubuko.com,布布扣bubuko.com,布布扣 is a basis for W that is also an orthogonal set.
  3. 一个向量在某一维的投影:ybubuko.com,布布扣^bubuko.com,布布扣=projbubuko.com,布布扣Lbubuko.com,布布扣y=y? ububuko.com,布布扣u?ububuko.com,布布扣bubuko.com,布布扣ububuko.com,布布扣 .
  4. An set is an orthonormal set if it is an orthogonal set of unit vectors.
  5. An m*n matrix U has orthonormal columns if and only if Ububuko.com,布布扣?bubuko.com,布布扣U=Ibubuko.com,布布扣
  6. 一个向量在某一空间的投影:ybubuko.com,布布扣^bubuko.com,布布扣=projbubuko.com,布布扣wbubuko.com,布布扣y=y? ububuko.com,布布扣1bubuko.com,布布扣bubuko.com,布布扣ububuko.com,布布扣1bubuko.com,布布扣?ububuko.com,布布扣1bubuko.com,布布扣bubuko.com,布布扣bubuko.com,布布扣ububuko.com,布布扣1bubuko.com,布布扣+y? ububuko.com,布布扣2bubuko.com,布布扣bubuko.com,布布扣ububuko.com,布布扣1bubuko.com,布布扣?ububuko.com,布布扣2bubuko.com,布布扣bubuko.com,布布扣bubuko.com,布布扣ububuko.com,布布扣2bubuko.com,布布扣 +...+y?ububuko.com,布布扣pbubuko.com,布布扣bubuko.com,布布扣ububuko.com,布布扣pbubuko.com,布布扣?ububuko.com,布布扣pbubuko.com,布布扣bubuko.com,布布扣bubuko.com,布布扣ububuko.com,布布扣pbubuko.com,布布扣.bubuko.com,布布扣
  7. 如何将一堆向量弄成正交单位向量: repeat 3.
  8. QR分解:如果A有线性无关的列向量,那么可以分解成Q(正交向量)和R(上三角矩阵,就是原坐标在正交坐标系的系数)Qbubuko.com,布布扣?bubuko.com,布布扣A=Qbubuko.com,布布扣?bubuko.com,布布扣(QR)=IR=Rbubuko.com,布布扣
  9. 最小平方lse(机器学习基础:非贝叶斯条件下的线性拟合问题),由Abubuko.com,布布扣?bubuko.com,布布扣(b?Axbubuko.com,布布扣^bubuko.com,布布扣)=0bubuko.com,布布扣 得到xbubuko.com,布布扣^bubuko.com,布布扣=(Abubuko.com,布布扣?bubuko.com,布布扣A)bubuko.com,布布扣?1bubuko.com,布布扣Abubuko.com,布布扣?bubuko.com,布布扣bbubuko.com,布布扣 。如果A可逆,此式可以化简。如果可以做QR分解,那么xbubuko.com,布布扣^bubuko.com,布布扣=Rbubuko.com,布布扣?1bubuko.com,布布扣Qbubuko.com,布布扣?bubuko.com,布布扣bbubuko.com,布布扣 .
  10. 函数内积的概念。

第七节:Diagonaliztion of Symmetric matrixs

  1. 如果一个矩阵是对称的,那么它的任何两个特征值所对应的特征空间是正交的。
  2. 矩阵可正交对角化等价于它是一个对称矩阵。
  3. A=PDPbubuko.com,布布扣?1bubuko.com,布布扣bubuko.com,布布扣 可以得到PCA(机器学习算法主成分分析,对协方差矩阵(对称)做对角化)
  4. 将二次方程转化成没有叉乘项的形式。x=Py,  A=PDPbubuko.com,布布扣?1bubuko.com,布布扣bubuko.com,布布扣 .
  5. 对于二次函数xbubuko.com,布布扣?bubuko.com,布布扣Axbubuko.com,布布扣 ,在|x| = 1的条件下,最大值为最大特征值,最小值为最小特征值。如果最大特征值(xbubuko.com,布布扣?bubuko.com,布布扣ububuko.com,布布扣1bubuko.com,布布扣bubuko.com,布布扣 )不能选,则选择次之。
  6. 正交矩阵P大概意思就是在该坐标系下,函数比较对称,D为坐标轴的伸展比例。
  7. SVD分解(该书的最后一个内容,蕴含了很多上述的内容)是要将矩阵分解成类似PDP^-1的形式,但是不是任何矩阵都能表示成这种形式(有n个线性无关的特征向量,正交的话还要是对称矩阵)。其中A=UΣVbubuko.com,布布扣?bubuko.com,布布扣bubuko.com,布布扣 Σbubuko.com,布布扣 是A的singular value(Abubuko.com,布布扣?bubuko.com,布布扣Abubuko.com,布布扣 的特征值的开方),V是Abubuko.com,布布扣?bubuko.com,布布扣Abubuko.com,布布扣 的对应特征向量,U是AVbubuko.com,布布扣 的归一化。AV内的向量是垂直的。UΣbubuko.com,布布扣 是AV的另外一种表示。

线性代数《Linear Algebra and Its Application》学习总结,布布扣,bubuko.com

线性代数《Linear Algebra and Its Application》学习总结

原文:http://www.cnblogs.com/wead-hsu/p/3712587.html

(0)
(0)
   
举报
评论 一句话评论(0
关于我们 - 联系我们 - 留言反馈 - 联系我们:wmxa8@hotmail.com
© 2014 bubuko.com 版权所有
打开技术之扣,分享程序人生!