Question: Let\(y = \left( {\begin{aligned}{}7\\9\end{aligned}} \right)\),\({{\bf{u}}_1} = \left( {\begin{aligned}{}{{1 \mathord{\left/

{\vphantom {1 {\sqrt {10} }}} \right.} {\sqrt {10} }}}\\{{{ - 3} \mathord{\left/{\vphantom {{ - 3} {\sqrt {10} }}} \right.} {\sqrt {10} }}}\end{aligned}} \right)\), and\(W = {\bf{Span}}\left\{ {{{\bf{u}}_1}} \right\}\).

  1. Let \(U\)be the\(2 \times 1\)matrix whose only column is\({{\bf{u}}_1}\). Compute\({U^T}U\)and\(U{U^T}\).
  2. Compute\({\bf{Pro}}{{\bf{j}}_W}y\)and\(\left( {{U^T}U} \right){\bf{y}}\).

Short Answer

Expert verified
    1. The matrices are\({U^T}U = \left( 1 \right)\),\(U{U^T} = \left( {\begin{aligned}{}{{1 \mathord{\left/{\vphantom {1 {10}}} \right.} {10}}}&{{{ - 3} \mathord{\left/{\vphantom {{ - 3} {10}}} \right.} {10}}}\\{{{ - 3} \mathord{\left/{\vphantom {{ - 3} {10}}} \right.} {10}}}&{{9 \mathord{\left/{\vphantom {9 {10}}} \right.} {10}}}\end{aligned}} \right)\).
    2. \({\rm{Pro}}{{\rm{j}}_W}{\bf{y}} = U{U^T}{\bf{y}} = \left( {\begin{aligned}{}{ - 2}\\6\end{aligned}} \right)\).

Step by step solution

01

Form \(U = \left( {{{\bf{u}}_1}} \right)\)

(a)

Given vector is\({{\bf{u}}_1} = \left( {\begin{aligned}{}{{1 \mathord{\left/

{\vphantom {1 {\sqrt {10} }}} \right.} {\sqrt {10} }}}\\{ - {3 \mathord{\left/

{\vphantom {3 {\sqrt {10} }}} \right.} {\sqrt {10} }}}\end{aligned}} \right)\).

Find\(U = \left( {{{\bf{u}}_1}} \right)\).

\(U = \left( {\begin{aligned}{}{{1 \mathord{\left/{\vphantom {1 {\sqrt {10} }}} \right.} {\sqrt {10} }}}\\{ - {3 \mathord{\left/{\vphantom {3 {\sqrt {10} }}} \right.} {\sqrt {10} }}}\end{aligned}} \right)\)

02

Determine \({U^T}U\) and \(U{U^T}\)

First, find the transpose of\(U\), that is\({U^T}\).

\({U^T} = \left( {\begin{aligned}{}{{1 \mathord{\left/

{\vphantom {1 {\sqrt {10} }}} \right.} {\sqrt {10} }}}&{ - {3 \mathord{\left/

{\vphantom {3 {\sqrt {10} }}} \right.} {\sqrt {10} }}}\end{aligned}} \right)\)

Now, compute\({U^T}U\)and\(U{U^T}\).

\(\begin{aligned}{c}{U^T}U = \left( {\begin{aligned}{}{{1 \mathord{\left/

{\vphantom {1 {\sqrt {10} }}} \right.} {\sqrt {10} }}}&{ - {3 \mathord{\left/

{\vphantom {3 {\sqrt {10} }}} \right.} {\sqrt {10} }}}\end{aligned}} \right)\left( {\begin{aligned}{}{{1 \mathord{\left/

{\vphantom {1 {\sqrt {10} }}} \right.} {\sqrt {10} }}}\\{ - {3 \mathord{\left/

{\vphantom {3 {\sqrt {10} }}} \right.} {\sqrt {10} }}}\end{aligned}} \right)\\ = \left( {\frac{1}{{10}} + \frac{9}{{10}}} \right)\\ = \left( 1 \right)\end{aligned}\)

And,

\(\begin{aligned}{c}U{U^T} = \left( {\begin{aligned}{}{{1 \mathord{\left/

{\vphantom {1 {\sqrt {10} }}} \right.

\kern-\nulldelimiterspace} {\sqrt {10} }}}\\{ - {3 \mathord{\left/

{\vphantom {3 {\sqrt {10} }}} \right.

\kern-\nulldelimiterspace} {\sqrt {10} }}}\end{aligned}} \right)\left( {\begin{aligned}{}{{1 \mathord{\left/

{\vphantom {1 {\sqrt {10} }}} \right.

\kern-\nulldelimiterspace} {\sqrt {10} }}}&{ - {3 \mathord{\left/

{\vphantom {3 {\sqrt {10} }}} \right.

\kern-\nulldelimiterspace} {\sqrt {10} }}}\end{aligned}} \right)\\ = \left( {\begin{aligned}{}{{1 \mathord{\left/

{\vphantom {1 {10}}} \right.

\kern-\nulldelimiterspace} {10}}}&{{{ - 3} \mathord{\left/

{\vphantom {{ - 3} {10}}} \right.

\kern-\nulldelimiterspace} {10}}}\\{{{ - 3} \mathord{\left/

{\vphantom {{ - 3} {10}}} \right.

\kern-\nulldelimiterspace} {10}}}&{{9 \mathord{\left/

{\vphantom {9 {10}}} \right.

\kern-\nulldelimiterspace} {10}}}\end{aligned}} \right)\end{aligned}\)

Hence,\({U^T}U = \left( 1 \right)\)and\(U{U^T} = \left( {\begin{aligned}{}{{1 \mathord{\left/

{\vphantom {1 {10}}} \right.

\kern-\nulldelimiterspace} {10}}}&{{{ - 3} \mathord{\left/

{\vphantom {{ - 3} {10}}} \right.

\kern-\nulldelimiterspace} {10}}}\\{{{ - 3} \mathord{\left/

{\vphantom {{ - 3} {10}}} \right.

\kern-\nulldelimiterspace} {10}}}&{{9 \mathord{\left/

{\vphantom {9 {10}}} \right.

\kern-\nulldelimiterspace} {10}}}\end{aligned}} \right)\).

03

Write the Theorem

Let \({\bf{y}} \in {\mathbb{R}^n}\), then \({\rm{Pro}}{{\rm{j}}_W}{\bf{y}} = U{U^T}{\bf{y}}\), if \(U = \left( {\begin{aligned}{}{{{\bf{u}}_1}}&{{{\bf{u}}_2}}& \cdots &{{{\bf{u}}_p}}\end{aligned}} \right)\).

04

Find \({\bf{Pro}}{{\bf{j}}_W}y\) and \(\left( {{U^T}U} \right)y\)

(b)

It is observed\(U = \left( {{{\bf{u}}_1}} \right)\)from part (a), then\({\rm{Pro}}{{\rm{j}}_W}{\bf{y}} = U{U^T}{\bf{y}}\).

From part (a), \(U{U^T} = \left( {\begin{aligned}{}{{1 \mathord{\left/{\vphantom {1 {10}}} \right.} {10}}}&{{{ - 3} \mathord{\left/{\vphantom {{ - 3} {10}}} \right.} {10}}}\\{{{ - 3} \mathord{\left/{\vphantom {{ - 3} {10}}} \right.} {10}}}&{{9 \mathord{\left/

{\vphantom {9 {10}}} \right} {10}}}\end{aligned}} \right)\), then find \(U{U^T}{\bf{y}}\), where \({\bf{y}} = \left( {\begin{aligned}{}7\\9\end{aligned}} \right)\).

\(\begin{aligned}{c}U{U^T}{\bf{y}} = \left( {\begin{aligned}{}{{1 \mathord{\left/

{\vphantom {1 {10}}} \right} {10}}}&{{{ - 3} \mathord{\left/{\vphantom {{ - 3} {10}}} \right.} {10}}}\\{{{ - 3} \mathord{\left/{\vphantom {{ - 3} {10}}} \right.} {10}}}&{{9 \mathord{\left/ {\vphantom {9 {10}}} \right.

\kern-\nulldelimiterspace} {10}}}\end{aligned}} \right)\left( {\begin{aligned}{}7\\9\end{aligned}} \right)\\ = \left( {\begin{aligned}{}{\frac{7}{{10}} - \frac{{27}}{{10}}}\\{ - \frac{{21}}{{10}} + \frac{{81}}{{10}}}\end{aligned}} \right)\\ = \left( {\begin{aligned}{}{ - \frac{{20}}{{10}}}\\{\frac{{60}}{{10}}}\end{aligned}} \right)\\ = \left( {\begin{aligned}{}{ - 2}\\6\end{aligned}} \right)\end{aligned}\)

So,\({\rm{Pro}}{{\rm{j}}_W}{\bf{y}} = \left( {\begin{aligned}{}{ - 2}\\6\end{aligned}} \right)\), as\({\rm{Pro}}{{\rm{j}}_W}{\bf{y}} = U{U^T}{\bf{y}}\).

Unlock Step-by-Step Solutions & Ace Your Exams!

  • Full Textbook Solutions

    Get detailed explanations and key concepts

  • Unlimited Al creation

    Al flashcards, explanations, exams and more...

  • Ads-free access

    To over 500 millions flashcards

  • Money-back guarantee

    We refund you if you fail your exam.

Over 30 million students worldwide already upgrade their learning with Vaia!

One App. One Place for Learning.

All the tools & learning materials you need for study success - in one app.

Get started for free

Most popular questions from this chapter

Let \({\mathbb{R}^{\bf{2}}}\) have the inner product of Example 1. Show that the Cauchy-Schwarz inequality holds for \({\bf{x}} = \left( {{\bf{3}}, - {\bf{2}}} \right)\) and \({\bf{y}} = \left( { - {\bf{2}},{\bf{1}}} \right)\). (Suggestion: Study \({\left| {\left\langle {{\bf{x}},{\bf{y}}} \right\rangle } \right|^{\bf{2}}}\).)

Use the Gram–Schmidt process as in Example 2 to produce an orthogonal basis for the column space of

\(A = \left( {\begin{aligned}{{}{r}}{ - 10}&{13}&7&{ - 11}\\2&1&{ - 5}&3\\{ - 6}&3&{13}&{ - 3}\\{16}&{ - 16}&{ - 2}&5\\2&1&{ - 5}&{ - 7}\end{aligned}} \right)\)

In Exercises 13 and 14, find the best approximation to\[{\bf{z}}\]by vectors of the form\[{c_1}{{\bf{v}}_1} + {c_2}{{\bf{v}}_2}\].

13.\[z = \left[ {\begin{aligned}3\\{ - 7}\\2\\3\end{aligned}} \right]\],\[{{\bf{v}}_1} = \left[ {\begin{aligned}2\\{ - 1}\\{ - 3}\\1\end{aligned}} \right]\],\[{{\bf{v}}_2} = \left[ {\begin{aligned}1\\1\\0\\{ - 1}\end{aligned}} \right]\]

In Exercises 1-4, find a least-sqaures solution of \(A{\bf{x}} = {\bf{b}}\) by (a) constructing a normal equations for \({\bf{\hat x}}\) and (b) solving for \({\bf{\hat x}}\).

4. \(A = \left( {\begin{aligned}{{}{}}{\bf{1}}&{\bf{3}}\\{\bf{1}}&{ - {\bf{1}}}\\{\bf{1}}&{\bf{1}}\end{aligned}} \right)\), \({\bf{b}} = \left( {\begin{aligned}{{}{}}{\bf{5}}\\{\bf{1}}\\{\bf{0}}\end{aligned}} \right)\)

Suppose the x-coordinates of the data \(\left( {{x_1},{y_1}} \right), \ldots ,\left( {{x_n},{y_n}} \right)\) are in mean deviation form, so that \(\sum {{x_i}} = 0\). Show that if \(X\) is the design matrix for the least-squares line in this case, then \({X^T}X\) is a diagonal matrix.

See all solutions

Recommended explanations on Math Textbooks

View all explanations

What do you think about this solution?

We value your feedback to improve our textbook solutions.

Study anywhere. Anytime. Across all devices.

Sign-up for free