N.B. Il formulario contiene alcune formule che non sono state trattate a lezione.

Si prega di ignorarle.
FORMULARIO DI STATISTICA DESCRITTIVA

𝑥max −𝑥min 𝑓𝑖 𝑛𝑖
𝑑= 𝑘 ℎ𝑖 = 𝑑𝑖 𝑜𝑝𝑝𝑢𝑟𝑒 ℎ𝑖 = 𝑑𝑖 𝑑𝑖 = 𝑥𝑖 − 𝑥𝑖−1

𝑥( 𝑛 ) +𝑥( 𝑛 +1)
⎧ 2 2
𝑠𝑒 𝑛 𝑒 ̀ 𝑝𝑎𝑟𝑖
{ 2
𝑀𝑒 = 𝑀 𝑒 ≅ 𝑥𝑀𝑒−1 + (𝑥𝑀𝑒 − 𝑥𝑀𝑒−1 ) 𝐹0,5−𝐹 𝑀𝑒−1
⎨ −𝐹 𝑀𝑒 𝑀𝑒−1
{ 𝑥( 𝑛2 +1) 𝑠𝑒 𝑛 𝑒 ̀ 𝑑𝑖𝑠𝑝𝑎𝑟𝑖
⎩

𝑥( 𝑛 ) +𝑥( 𝑛 +1)
⎧ 4 4
𝑠𝑒 𝑛 𝑒 ̀ 𝑝𝑎𝑟𝑖
{ 2 0,25−𝐹(𝑄1 −1)
𝑄1 = ⎨ 𝑄1 ≅ 𝑥𝑄1 −1 + (𝑥𝑄1 − 𝑥𝑄1 −1 ) 𝐹
(𝑄1 ) −𝐹(𝑄1 −1)
{ 𝑥( 𝑛+1 ) 𝑠𝑒 𝑛 𝑒 ̀ 𝑑𝑖𝑠𝑝𝑎𝑟𝑖
⎩ 4

𝑥( 3𝑛 ) +𝑥( 3𝑛 +1 )
⎧ 4 4
𝑠𝑒 𝑛 𝑒 ̀ 𝑝𝑎𝑟𝑖
{ 2 0,75−𝐹(𝑄3 −1)
𝑄3 = ⎨ 𝑄3 ≅ 𝑥𝑄3 −1 + (𝑥𝑄3 − 𝑥𝑄3 −1 ) 𝐹
(𝑄3 ) −𝐹(𝑄3 −1)
{ 𝑥( 3𝑛 +1) 𝑠𝑒 𝑛 𝑒 ̀ 𝑑𝑖𝑠𝑝𝑎𝑟𝑖
⎩ 4

𝜇𝑥 = 𝑥 = 𝑀 (𝑥) = 1
𝑛 ∑𝑛𝑖=1 𝑥𝑖 𝑛𝑖 𝜇2 = 𝑀 (𝑥2 ) = 1
𝑛 ∑𝑛𝑖=1 𝑥2𝑖 𝑛𝑖

𝜎𝑥2 = 1
𝑛 ∑𝑛𝑖=1 (𝑥𝑖 − 𝑥)2 𝑛𝑖 = 𝑀 (𝑥2 ) − (𝑥)2 𝜎𝑥 = √𝜎𝑥2 = √ 𝑛1 ∑𝑛𝑖=1 (𝑥𝑖 − 𝑥)2 𝑛𝑖

𝜎𝑥
𝑅𝑎𝑛𝑔𝑒 (𝑥) = 𝑥(𝑛) − 𝑥(1) 𝐼𝑄𝑅 = 𝑄3 − 𝑄1 𝐶𝑉 = 𝜇𝑥

𝐻 = ∑𝑖=1 𝑓𝑖 log (𝑓𝑖 )

𝑘 𝑘 𝑘 𝐻
𝐺 = 1 − ∑𝑖=1 𝑓𝑖2 𝐺∗ = 𝐺 𝑘−1 𝐻∗ = log(𝑘)

1
∑𝑛 (𝑥𝑖 −𝑥)3 𝑛𝑖 3 4
𝐴𝐹 = 𝑛 𝑖=1
3
𝜎𝑥 = 1
𝑛 ∑𝑛𝑖=1 ( 𝑥𝜎𝑖 −𝑥 ) 𝑛𝑖 𝐾= 1
𝑛 ∑𝑛𝑖=1 ( 𝑥𝜎𝑖 −𝑥 ) 𝑛𝑖 − 3
𝑥 𝑥

2
(𝑛𝑖𝑗 −̂
𝑛𝑖𝑗 ) 𝜒2
𝜒2 = ∑𝑘𝑖=1 ∑ℎ𝑗=1 ̂
𝑛𝑖𝑗 𝜙2 = 𝑛

𝜎𝑥𝑦 = 1
𝑛 ∑𝑛𝑖=1 (𝑥𝑖 − 𝜇𝑥 ) (𝑦𝑖 − 𝜇𝑦 ) 𝑛𝑖 = 𝑀 (𝑥𝑦) − 𝑀 (𝑥) 𝑀 (𝑦)
1 𝑛
𝜎𝑥𝑦 𝑛 ∑𝑖=1 𝑥𝑖 𝑦𝑖 −𝜇𝑥 𝜇𝑦
𝜌𝑥𝑦 = 𝜎𝑥 𝜎𝑦 = 𝑛
1
√𝑛 ∑𝑖=1 (𝑥𝑖 −𝑥)2 𝑛𝑖 √ 𝑛
1
∑𝑛 (𝑦 −𝑦)2 𝑛𝑖
𝑖=1 𝑖

∑𝑛−1 (𝑝𝑖 −𝑞𝑖 )

𝑅= 𝑖=1 ̃ = 1 − ∑𝑛−1 (𝑝𝑖+1 − 𝑝𝑖 ) (𝑞𝑖+1 − 𝑞𝑖 )
𝑅
∑𝑛−1 𝑝
𝑖=1 𝑖
𝑖=0

𝑦 ̂ = 𝛽0̂ + 𝛽1̂ 𝑥 𝛽1̂ = 𝐶𝑜𝑣(𝑥,𝑦)

𝜎𝑥2 𝛽0̂ = 𝑦 − 𝛽1̂ 𝑥

1
 _________________________________________________________

FORMULARIO DI STATISTICA INFERENZIALE

𝑃 (∅) = 0

𝑃 (𝐴) = 1 − 𝑃 (𝐴)

𝑃 (𝐴 ∪ 𝐵) = 𝑃 (𝐴) + 𝑃 (𝐵) − 𝑃 (𝐴 ∩ 𝐵)

𝑃(𝐴∩𝐵) 𝑃(𝐵∩𝐴)
𝑃 (𝐴|𝐵) = 𝑃(𝐵) 𝑃 (𝐵) > 0 𝑃 (𝐵|𝐴) = 𝑃(𝐴) 𝑃 (𝐴) > 0

𝑃 (𝐴 ∩ 𝐵) = 𝑃 (𝐵) ⋅ 𝑃 (𝐴|𝐵) = 𝑃 (𝐴) ⋅ 𝑃 (𝐵|𝐴) 𝑃 (𝐴 ∩ 𝐵) = 𝑃 (𝐴) ⋅ 𝑃 (𝐵)

𝑃(𝐻𝑖 ∩𝐸) 𝑃(𝐻𝑖 )⋅𝑃(𝐸|𝐻𝑖 )

𝑃 (𝐻𝑖 |𝐸) = 𝑃(𝐸) = ∑𝑚 𝑃(𝐻𝑗 )⋅𝑃(𝐸|𝐻𝑗 )
𝑗=1

+∞
𝜇𝑟 = 𝐸 (𝑋 𝑟 ) = ∑ 𝑥𝑟𝑖 𝑝𝑖 𝜇𝑟 = 𝐸 (𝑋 𝑟 ) = ∫ 𝑥𝑟 𝑓 (𝑥) 𝑑𝑥
−∞

+∞
𝜇 = 𝐸 (𝑋) = ∑ 𝑥𝑖 𝑝𝑖 𝜇 = 𝐸 (𝑋) = ∫ 𝑥𝑓 (𝑥) 𝑑𝑥
−∞

+∞
𝑉 𝐴𝑅 (𝑋) = 𝐸 (𝑋 − 𝜇)2 = ∑ (𝑥𝑖 − 𝜇)2 𝑝𝑖 = 𝐸 (𝑋 2 ) − [𝐸 (𝑋)]2 𝑉 𝐴𝑅 (𝑋) = ∫ (𝑥𝑖 − 𝜇)2 𝑓 (𝑥) 𝑑𝑥
−∞

1 𝑏+𝑎 (𝑏−𝑎)2
𝑋 ∼ 𝑈 (𝑎, 𝑏) 𝑃 (𝑋 = 𝑥) = 𝑏−𝑎 (cost.) 𝐸 (𝑋) = 2 𝑉 𝐴𝑅 (𝑋) = 12

𝑋 ∼ 𝐵𝑒𝑟 (𝑝) 𝑃 (𝑋 = 𝑥) = 𝑝𝑥 (1 − 𝑝)1−𝑥 𝐸 (𝑋) = 𝑝 𝑉 𝐴𝑅 (𝑋) = 𝑝 (1 − 𝑝)

𝑛
𝑋 ∼ 𝐵𝑖𝑛 (𝑛 , 𝑝) 𝑃 (𝑋 = 𝑥) = ( ) 𝑝𝑥 (1 − 𝑝)𝑛−𝑥 𝐸 (𝑋) = 𝑛𝑝 𝑉 𝐴𝑅 (𝑋) = 𝑛𝑝 (1 − 𝑝)
𝑥

⎛ 𝑆 ⎞ 𝑁 −𝑆
⎜ ⎟⎛⎜ ⎞
⎟
⎝ 𝑥 ⎠⎝ 𝑛−𝑥 ⎠ 𝑆
𝑋 ∼ 𝐻𝑔 (𝑁 , 𝑆, 𝑛) 𝑃 (𝑋 = 𝑥) = 𝐸 (𝑋) = 𝑛𝑝 (𝑝 = 𝑁) 𝑉 𝐴𝑅 (𝑋) = 𝑛𝑝 (1 − 𝑝) ( 𝑁−𝑛
𝑁−1 )
⎛ 𝑁 ⎞
⎜ ⎟
⎝ 𝑛 ⎠

𝑒−𝜆
𝑋 ∼ 𝑃 𝑜 (𝜆) 𝑃 (𝑋 = 𝑥) = 𝜆𝑥 ⋅ 𝑥! 𝐸 (𝑋) = 𝜆 𝑉 𝐴𝑅 (𝑋) = 𝜆

𝑋 ∼ 𝐺𝑒𝑜 (𝑝) 𝑃 (𝑋 = 𝑥) = 𝑝 (1 − 𝑝)𝑥−1 𝐸 (𝑋) = 1

𝑝 𝑉 𝐴𝑅 (𝑋) = 1−𝑝
𝑝2

2
 1 𝑥−𝜇 2
𝑋 ∼ 𝑁 (𝜇 , 𝜎2 ) 𝑓 (𝑥) = √ 1 𝑒− 2 ( 𝜎 ) 𝐸 (𝑋) = 𝜇 𝑉 𝐴𝑅 (𝑋) = 𝜎2
2𝜋𝜎2

𝑥−𝜇
𝑍= 𝜎

𝑋 ∼ 𝜒2(𝑟) 𝐸 (𝑋) = 𝑟 𝑉 𝐴𝑅 (𝑋) = 2𝑟

𝑟
𝑋 ∼ 𝑡(𝑟) 𝐸 (𝑋) = 0 𝑉 𝐴𝑅 (𝑋) = 𝑟−2

𝑉1 /𝑟1 𝜒2(𝑟 ) /𝑟1

𝑋 ∼ 𝐹(𝑟1 , 𝑟2 ) 𝑉2 /𝑟2 = 1
𝜒2(𝑟 ) /𝑟2
∼ 𝐹(𝑟1 ,𝑟2 )
2

𝑆𝑛 𝑆𝑛
𝑛 ∼ 𝐵𝑖𝑛 (𝑛 , 𝑝) 𝑛 ∼ 𝑁 (𝑝, 𝑝(1−𝑝)
𝑛 ) 𝑝𝑒𝑟 𝑛 𝑔𝑟𝑎𝑛𝑑𝑒

𝜎2
𝑋 ∼ 𝑁 (𝜇 , 𝜎2 ) 𝐸 (𝑋) = 𝜇 𝑉 𝐴𝑅 (𝑋) = 𝑛

1 2 (𝑛−1)𝑆 2 𝑋−𝜇
𝑆2 = 𝑛−1 ∑ (𝑥𝑖 − 𝑋) 𝜎2 ∼ 𝜒2(𝑛−1) 𝑇 = √𝑆
∼ 𝑡(𝑛−1)
𝑛

𝜎12 𝜎22
(𝑋1 − 𝑋2 ) ∼ 𝑁 (𝜇1 − 𝜇2 , 𝑛1 + 𝑛2 )

(𝑋1 −𝑋2 )−(𝜇1 −𝜇2 )

𝜎12 , 𝜎22 𝑛𝑜𝑡𝑒 𝑍= 2 2
∼ 𝑁 (0, 1)
√ 𝜎1 + 𝜎2
𝑛1 𝑛2

(𝑋1 −𝑋2 )−(𝜇1 −𝜇2 ) (𝑛1 −1)𝑆12 +(𝑛2 −1)𝑆22

𝜎12 = 𝜎22 𝑖𝑛𝑐𝑜𝑔𝑛𝑖𝑡𝑒 𝑇 = (𝑛 −1)𝑆2 +(𝑛2 −1)𝑆2
𝑆𝑝2 = 𝑛1 +𝑛2 −2 ∼ 𝑡(𝑛1 +𝑛2 −2)
√( 1 𝑛 1+𝑛 −2 2 )( 1 + 1 )
𝑛1 𝑛2
1 2
_________________________________________________________

INTERVALLI DI CONFIDENZA

𝑋 ∼ 𝑁 (𝜇 = ? , 𝜎2 ) 𝑃 [𝑥 − 𝑧 𝛼2 √𝜎𝑛 ≤ 𝜇 ≤ 𝑥 + 𝑧 𝛼2 √𝜎𝑛 ] = (1 − 𝛼)

𝑋 ∼ 𝑁 (𝜇 = ? , 𝜎2 = ?) 𝑃 [𝑥 − 𝑡 𝛼2 √𝑆𝑛 ≤ 𝜇 ≤ 𝑥 + 𝑡 𝛼2 √𝑆𝑛 ] = (1 − 𝛼)

2
(𝑛−1)𝑆 2
𝑋 ∼ 𝑁 (𝜇 , 𝜎2 = ?) 𝑃 [ 𝜒(𝑛−1)𝑆
2 ,𝛼
≤ 𝜎2 ≤ 𝜒2(𝑛−1) ,(1− 𝛼 )
] = (1 − 𝛼)
(𝑛−1) 2 2

𝑆𝑛 𝑆𝑛 𝑆𝑛 𝑆𝑛
𝑛 (1− 𝑛 ) 𝑛 (1− 𝑛 )
𝑆𝑛
𝑛 ∼ 𝑁 (𝑝 = ?, 𝑝(1−𝑝)
𝑛 = ?) 𝑃 [ 𝑆𝑛𝑛 − 𝑧 𝛼2 √ 𝑛 ≤𝑝≤ 𝑆𝑛
𝑛 + 𝑧 𝛼2 √ 𝑛 ] = (1 − 𝛼)

𝑋 ∼ 𝑁 (𝜇1 = ? , 𝜎12 ) 𝑋 ∼ 𝑁 (𝜇2 = ? , 𝜎22 )

𝜎2 𝜎22 𝜎2 𝜎22
𝑃 [(𝑋1 − 𝑋2 ) − 𝑧 𝛼2 √ 𝑛1 + 𝑛2 ≤ (𝜇1 − 𝜇2 ) ≤ (𝑋1 − 𝑋2 ) + 𝑧 𝛼2 √ 𝑛1 + 𝑛2 ] = (1 − 𝛼)
1 1

𝑋 ∼ 𝑁 (𝜇1 , 𝜎12 = ?) 𝑋 ∼ 𝑁 (𝜇2 , 𝜎22 = ?) 𝜎12 = 𝜎22 = ?

𝑃 [(𝑋1 − 𝑋2 ) − 𝑡𝑡(𝑛 𝛼
𝑆𝑝 √ 𝑛1 + 1
𝑛2 ≤ (𝜇1 − 𝜇2 ) ≤ (𝑋1 − 𝑋2 ) − 𝑡𝑡(𝑛 𝛼
𝑆𝑝 √ 𝑛1 + 1
𝑛2 ] = (1 − 𝛼)
1 +𝑛2 −2, 2 ) 1 1 +𝑛2 −2, 2 ) 1

(𝑛1 −1)𝑆12 +(𝑛2 −1)𝑆22

𝑆𝑝2 = 𝑛1 +𝑛2 −2

3
_________________________________________________________

TEST DELLE IPOTESI

𝑋−𝜇 𝑋−𝜇
𝑍= √𝜎
𝑇 = √𝑠
𝑛 𝑛

(𝑋1 −𝑋2 )−(𝜇1 −𝜇2 ) (𝑋1 −𝑋2 )−(𝜇1 −𝜇2 )

𝑍= 2 2
𝑇 = (𝑛1 −1)𝑆2 2
√ 𝜎1 + 𝜎2 √( 1 +(𝑛2 −1)𝑆2 )( 1 + 1 )
𝑛1 𝑛2 𝑛1 +𝑛2 −2 𝑛1 𝑛2

𝑆𝑛 𝑆𝑛2
( 1 − 𝑛 )−(𝑝1 −𝑝2 )
𝑆𝑛
𝑛 −𝑝0 𝑛
𝑍= ≅ 𝑁 (0, 1) 𝑍= √ 𝑆𝑛 𝑆𝑛 𝑆𝑛 𝑆𝑛
√ 𝑝0 (1−𝑝0 ) √
√ 1 1
𝑛 (1− 𝑛 )
2 2
𝑛 (1− 𝑛 )
𝑛
+
⎷ 𝑛1 𝑛2

(𝑛−1)𝑆 2
𝜎2 ≅ 𝜒2(𝑛−1)

2 2
𝑘
(𝑓𝑖 − 𝑝𝑖 )2 𝑘
(𝑛 − 𝑛𝑝𝑖 )2 𝑘 𝑘
(𝑛𝑖𝑗 − 𝑐𝑖𝑗 ) 𝑘 𝑘
(𝑛𝑖𝑗 − 𝑛̂ 𝑖𝑗 )
𝜒2 = ∑ =∑ 𝑖 𝜒2 = ∑ ∑ = ∑∑
𝑖=1
𝑝𝑖 𝑖=1
𝑛𝑝𝑖 𝑖=1 𝑗=1
𝑐𝑖𝑗 𝑖=1 𝑗=1
𝑛̂ 𝑖𝑗

̌ 𝑠𝑒𝑣∶
𝐷𝑖𝑠𝑢𝑔𝑢𝑎𝑔𝑙𝑖𝑎𝑛𝑧𝑎 𝑑𝑖 𝐶𝑒𝑏𝑦 ̌ ̈ 𝑃 {|𝑋 − 𝜇| < 𝜀𝜎} ≥ 1 − 1
𝑃 {|𝑋 − 𝜇| < 𝑘} ≥ 1 − 𝜎2
𝜀2 𝑘2

REGRESSIONE

1 𝑛 1 𝑛
𝑥= 𝑛 ∑𝑖=1 𝑥𝑖 𝑦= 𝑛 ∑𝑖=1 𝑦𝑖

(𝑥𝑖 −𝑥)2 (𝑦𝑖 −𝑦)2

𝐶𝑜𝑣 (𝑋, 𝑌) = ∑ 𝑥𝑖 𝑦𝑖 − (𝑥 𝑦) 𝑠2𝑥 = ∑𝑛𝑖=1 𝑛−1 𝑠2𝑦 = ∑𝑛𝑖=1 𝑛−1

𝐶𝑜𝑣(𝑥,𝑦)
𝑦 ̂ = 𝛽0∗ + 𝛽1∗ 𝑥 𝛽1∗ = 𝜎𝑥2 𝛽0∗ = 𝑦 − 𝛽1∗ 𝑥

(𝑛−1)𝑠2𝑦 − 𝐶𝑂𝑉(𝑋,𝑌)
𝑠𝜀 = √ 𝑆𝑆𝐸 √ 2 𝑠𝑥
𝑛−2 = 𝑛−2

𝛽1∗ −𝛽1 𝑠𝜀
𝑆𝑡𝑎𝑡𝑖𝑠𝑡𝑖𝑐𝑎 𝑡𝑒𝑠𝑡 ∶ 𝑇 = 𝑠𝛽∗ 𝑠𝛽1∗ = √(𝑛−1)𝑠2𝑥
1

𝑆𝑆𝐸
𝐼𝑛𝑡𝑒𝑟𝑣𝑎𝑙𝑙𝑜 𝑑𝑖 𝑐𝑜𝑛𝑓𝑖𝑑𝑒𝑛𝑧𝑎 ∶ 𝛽1∗ ± 𝑡 𝛼2 ,(𝑛−2) ⋅ 𝑠𝛽1∗ 𝑅2 = 1 − ∑(𝑦𝑖 −𝑦)2