Foundations — Rank of a matrix — definition, row rank = column rank theorem

4.5.12 · D1 · Maths › Linear Algebra (Full) › Rank of a matrix — definition, row rank = column rank theore

Parent theorem ("row rank = column rank") par trust karne se pehle, tumhe us chhoti vocabulary mein fluent hona chahiye jo wo silently assume karta hai. Ye page har word ko scratch se banata hai, us order mein jo agla word meaningful banaye.

0. Shuruaati picture: numbers as arrows

Ye picture. 2D mein, $(3,1)$ matlab "3 right chalo, 1 upar." Arrow hi vector hai.

Topic ko iske zaroort kyun hai. Parent note rows aur columns ko "vectors" kehta hai aur poochta hai ki wo kitni directions cover karte hain. Jab tak tum har row aur column ko ek arrow ki tarah nahi dekhte, tab tak tum directions count nahi kar sakte.

1. Matrix — ek grid, aur use padhne ke do tarike

$A=\left(\begin{array}{ccc}1& 2& 1\\ 2& 4& 3\\ 3& 6& 4\end{array}\right)\text{is }3\times 3.$

Usi grid ko padhne ke do tarike — yahi sab kuch ka core hai:

• Row reading: teen horizontal arrows, har ek ${\mathbb{R}}^3$ mein: $(1,2,1)$, $(2,4,3)$, $(3,6,4)$.
• Column reading: teen vertical arrows, har ek bhi ${\mathbb{R}}^3$ mein: $\left(\genfrac{}{}{0ex}{}{1}{2}\right)3$, $\left(\genfrac{}{}{0ex}{}{2}{4}\right)6$, $\left(\genfrac{}{}{0ex}{}{1}{3}\right)4$.

Topic ko dono readings ki zaroorat kyun hai. "Row rank" row arrows mein directions count karta hai; "column rank" column arrows mein directions count karta hai. Pura miracle yeh hai ki arrows ke ye do bilkul alag bundles same count dete hain.

Ye picture. Transpose = matrix ko side par tilt karna. $A$ ki rows $A^T$ ke columns ban jaati hain.

Topic ko iske zaroorat kyun hai. Proof ka Step 4 kehta hai "argument ko $A^T$ par apply karo." Yahi woh trick hai jo rows aur columns ko swap karke equality laati hai — isliye $A^T$ pehle bilkul clear hona chahiye.

2. Arrows ko scale aur add karna → linear combination

Ye picture. Mujhe arrows $v_1,v_2$ do aur koi bhi dials $s_1,s_2$; reachable tips poora ek region sweep karte hain. Woh region agle section ka star hai.

Topic ko iske zaroorat kyun hai. Parent note mein har ek move ek linear combination hai: row operations ("$R_2\to R_2-2R_1$"), factorization $A=CR$, "column 2 = 2×column 1." Agar linear combination fuzzy hai, to proof fog hai.

Matrix Multiplication as Linear Combination dekhein kyun $Ax$ khud bhi bas $A$ ke columns ki ek linear combination hai.

3. Span — har jagah jo tum reach kar sako

Ye picture, case by case (hum sab cover karte hain):

• Ek nonzero arrow ka span = usse hone wali infinite line.
• Do arrows alag directions mein ka span = poora plane jisme wo hai.
• Do arrows usi line par (ek doosre ka multiple hai) ka span = phir bhi bas ek line — doosre ne kuch add nahi kiya.
• Zero vector akele ka span = bas ek single point $(0,0)$.

Topic ko iske zaroorat kyun hai. "Row space = span of the rows" aur "column space = span of the columns." Rank measure karta hai ki ye spans kitne bade hain — ek point, ek line, ek plane, ek volume. Section 5 "kitna bada" ko ek number banata hai.

4. Linear independence — koi arrow doosron ki copy nahi

Ye pictures (sab cases):

• $(1,2)$ aur $(2,4)$: doosra $2\times$ pehle ka hai → dependent (same line). Yahi parent ka rank-1 trap hai.
• $(1,0)$ aur $(0,1)$: na hi ek doosre ka multiple hai → independent (wo poora plane khol dete hain).
• Koi bhi set jisme zero vector shamil ho → automatically dependent ($1\cdot \vec{0}$ plus zero dials deta hai $\vec{0}$).

Topic ko iske zaroorat kyun hai. Rank sirf independent directions count karta hai. Parent ka mistake box ("nonzero rows still be dependent ho sakti hain") bilkul ek warning hai ki independence, na ki "nonzero-ness," matter karta hai.

Aur Linear Independence and Basis mein.

5. Basis aur dimension — honest count

Topic ko iske zaroorat kyun hai. Ab "row rank = dim(row space)" aur "column rank = dim(column space)" fully defined hain: har ek us bundle ko span karne ke liye zaroori independent arrows ki sankhya hai. Rank ek dimension hai, plain and simple.

Linear Independence and Basis aur Row Space and Column Space dekhein.

6. Pivots — computational shortcut

Ye picture. Leading entries ki ek descending staircase; staircase ke neeche aur left mein sab zeros. Steps ki sankhya = pivots ki sankhya.

Topic ko iske zaroorat kyun hai. Pivots count karna rank paane ka practical tarika hai bina independence eyeballing ke. Parent ka rank ka teesra face — "number of pivots" — yahaan rehta hai. Full machinery: Gaussian Elimination & Echelon Form.

7. Kernel aur nullity — jo bach gaya (bound ke liye)

Ye picture. Wo directions jise transformation flatten kar deta hai — wo output tak survive nahi karte.

Topic ko iske zaroorat kyun hai. Parent ka Rank–Nullity line, $\mathrm{rank}(A)+\dim \ker (A)=n$, "jo directions survive karte hain" ko "jo directions crush ho jaate hain" se balance karta hai. Details: Rank–Nullity Theorem.

Ye foundations topic ko kaise feed karte hain

Equipment checklist

Connections

• Parent: Rank of a Matrix
• Row Space and Column Space
• Linear Independence and Basis
• Matrix Multiplication as Linear Combination
• Gaussian Elimination & Echelon Form
• Rank–Nullity Theorem
• Invertible Matrix Theorem