Subjects statistics

Covariance Calculation 36Be38

Step-by-step solutions with LaTeX - clean, fast, and student-friendly.

Search Solutions

Covariance Calculation 36Be38


1. **Énoncé du problème :** Nous avons une série double (x, y) avec des classes d'ancienneté $x$ et des classes de salaire $y$. Les centres des classes sont : - Ancienneté $x_i$ : $1, 3, 5, 8, 12.5$ (milieu des intervalles) - Salaire $y_j$ : $350, 450, 550, 750, 1200$ Les effectifs $n_{ij}$ sont donnés dans le tableau. 2. **Distribution des effectifs marginaux de $x$ :** Calculer $n_{i.} = \sum_j n_{ij}$ pour chaque $i$. 3. **Calcul de $\bar{x}$ et $\sigma_x$ :** Formules : $$\bar{x} = \frac{1}{N} \sum_i n_{i.} x_i$$ $$\sigma_x = \sqrt{\frac{1}{N} \sum_i n_{i.} (x_i - \bar{x})^2}$$ avec $N = \sum_i n_{i.}$ 4. **Distribution des effectifs marginaux de $y$ :** Calculer $n_{.j} = \sum_i n_{ij}$ pour chaque $j$. 5. **Calcul de $\bar{y}$ et $\sigma_y$ :** $$\bar{y} = \frac{1}{N} \sum_j n_{.j} y_j$$ $$\sigma_y = \sqrt{\frac{1}{N} \sum_j n_{.j} (y_j - \bar{y})^2}$$ 6. **Calcul de la covariance $\mathrm{cov}(x,y)$ :** $$\mathrm{cov}(x,y) = \frac{1}{N} \sum_i \sum_j n_{ij} (x_i - \bar{x})(y_j - \bar{y})$$ --- **Calculs détaillés :** - Effectifs $n_{ij}$ : \begin{align*} n_{11} &= 5, n_{12} = 3, n_{13} = 5, n_{14} = 2, n_{15} = 0 \\ n_{21} &= 2, n_{22} = 4, n_{23} = 2, n_{24} = 1, n_{25} = 0 \\ n_{31} &= 0, n_{32} = 0, n_{33} = 3, n_{34} = 2, n_{35} = 1 \\ n_{41} &= 0, n_{42} = 0, n_{43} = 1, n_{44} = 1, n_{45} = 4 \\ n_{51} &= 0, n_{52} = 1, n_{53} = 4, n_{54} = 0, n_{55} = 0 \end{align*} - Centres des classes : $$x = [1, 3, 5, 8, 12.5], \quad y = [350, 450, 550, 750, 1200]$$ - Effectifs marginaux $n_{i.}$ : \begin{align*} n_{1.} &= 5+3+5+2+0 = 15 \\ n_{2.} &= 2+4+2+1+0 = 9 \\ n_{3.} &= 0+0+3+2+1 = 6 \\ n_{4.} &= 0+0+1+1+4 = 6 \\ n_{5.} &= 0+1+4+0+0 = 5 \end{align*} - Total $N = 15 + 9 + 6 + 6 + 5 = 41$ - Moyenne $\bar{x}$ : $$\bar{x} = \frac{1}{41}(15\times1 + 9\times3 + 6\times5 + 6\times8 + 5\times12.5) = \frac{15 + 27 + 30 + 48 + 62.5}{41} = \frac{182.5}{41} \approx 4.451$$ - Variance $\sigma_x^2$ : \begin{align*} \sum_i n_{i.} (x_i - \bar{x})^2 &= 15(1-4.451)^2 + 9(3-4.451)^2 + 6(5-4.451)^2 + 6(8-4.451)^2 + 5(12.5-4.451)^2 \\ &= 15(11.906) + 9(2.104) + 6(0.302) + 6(12.566) + 5(64.086) \\ &= 178.59 + 18.94 + 1.81 + 75.40 + 320.43 = 595.18 \end{align*} $$\sigma_x = \sqrt{\frac{595.18}{41}} = \sqrt{14.52} \approx 3.81$$ - Effectifs marginaux $n_{.j}$ : \begin{align*} n_{.1} &= 5+2+0+0+0 = 7 \\ n_{.2} &= 3+4+0+0+1 = 8 \\ n_{.3} &= 5+2+3+1+4 = 15 \\ n_{.4} &= 2+1+2+1+0 = 6 \\ n_{.5} &= 0+0+1+4+0 = 5 \end{align*} - Moyenne $\bar{y}$ : $$\bar{y} = \frac{1}{41}(7\times350 + 8\times450 + 15\times550 + 6\times750 + 5\times1200) = \frac{2450 + 3600 + 8250 + 4500 + 6000}{41} = \frac{24800}{41} \approx 604.88$$ - Variance $\sigma_y^2$ : \begin{align*} \sum_j n_{.j} (y_j - \bar{y})^2 &= 7(350-604.88)^2 + 8(450-604.88)^2 + 15(550-604.88)^2 + 6(750-604.88)^2 + 5(1200-604.88)^2 \\ &= 7(65070) + 8(24000) + 15(3025) + 6(21000) + 5(353000) \\ &= 455490 + 192000 + 45375 + 126000 + 1765000 = 2586865 \end{align*} $$\sigma_y = \sqrt{\frac{2586865}{41}} = \sqrt{63119} \approx 251.23$$ - Covariance $\mathrm{cov}(x,y)$ : Calculer $$\sum_i \sum_j n_{ij} (x_i - \bar{x})(y_j - \bar{y})$$ Calcul intermédiaire par classes : \begin{align*} \text{Pour } i=1, x_1=1, x_1-\bar{x} = -3.451: \\ &5(350-604.88)(-3.451) + 3(450-604.88)(-3.451) + 5(550-604.88)(-3.451) + 2(750-604.88)(-3.451) + 0 = \\ &5(-254.88)(-3.451) + 3(-154.88)(-3.451) + 5(-54.88)(-3.451) + 2(145.12)(-3.451) = \\ &4397.5 + 1605.3 + 946.0 - 1001.9 = 6946.9 \\ \text{Pour } i=2, x_2=3, x_2-\bar{x} = -1.451: \\ &2(350-604.88)(-1.451) + 4(450-604.88)(-1.451) + 2(550-604.88)(-1.451) + 1(750-604.88)(-1.451) + 0 = \\ &2(-254.88)(-1.451) + 4(-154.88)(-1.451) + 2(-54.88)(-1.451) + 1(145.12)(-1.451) = \\ &739.7 + 899.3 + 159.4 - 210.5 = 1587.9 \\ \text{Pour } i=3, x_3=5, x_3-\bar{x} = 0.549: \\ &0 + 0 + 3(550-604.88)(0.549) + 2(750-604.88)(0.549) + 1(1200-604.88)(0.549) = \\ &3(-54.88)(0.549) + 2(145.12)(0.549) + 1(595.12)(0.549) = \\ &-90.4 + 159.4 + 326.8 = 395.8 \\ \text{Pour } i=4, x_4=8, x_4-\bar{x} = 3.549: \\ &0 + 0 + 1(550-604.88)(3.549) + 1(750-604.88)(3.549) + 4(1200-604.88)(3.549) = \\ &1(-54.88)(3.549) + 1(145.12)(3.549) + 4(595.12)(3.549) = \\ &-194.8 + 514.9 + 8457.3 = 8777.4 \\ \text{Pour } i=5, x_5=12.5, x_5-\bar{x} = 8.049: \\ &0 + 1(450-604.88)(8.049) + 4(550-604.88)(8.049) + 0 + 0 = \\ &1(-154.88)(8.049) + 4(-54.88)(8.049) = \\ &-1246.7 - 1767.3 = -3014.0 \end{align*} Somme totale : $$6946.9 + 1587.9 + 395.8 + 8777.4 - 3014.0 = 9693.0$$ Donc $$\mathrm{cov}(x,y) = \frac{9693.0}{41} \approx 236.4$$ --- **Réponse finale :** - $\bar{x} \approx 4.451$, $\sigma_x \approx 3.81$ - $\bar{y} \approx 604.88$, $\sigma_y \approx 251.23$ - $\mathrm{cov}(x,y) \approx 236.4$ --- **Desmos:** {"latex":"y=0","features":{"intercepts":true,"extrema":true}}