|
|
Для Герасимова 2. Разработка модели потребительских предпочтений на основе данных рекомендательной сети epinions com
Приложение В. Результаты анализа для случаев 6-9 кластеров
6 кластеров
Рисунок 0.36 – Диаграмма центра кластера 1 для случая 6 кластеров
[1] "Points: 103"
Call:
lm(formula = overall
durability + eou + battery + lag + picture,
data = x)
Residuals:
Min 1Q Median 3Q Max
-3.4330 -0.3231 0.0365 0.4575 1.7011
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) -1.23129 0.58474 -2.106 0.0378 *
durability 0.51757 0.08751 5.914 4.99e-08 ***
eou 0.20671 0.10967 1.885 0.0624 .
battery 0.13148 0.07269 1.809 0.0736 .
lag -0.01553 0.09050 -0.172 0.8641
picture 0.49337 0.08542 5.776 9.23e-08 ***
---
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
Residual standard error: 0.7715 on 97 degrees of freedom
(56 observations deleted due to missingness)
Multiple R-squared: 0.5621, Adjusted R-squared: 0.5396
F-statistic: 24.9 on 5 and 97 DF, p-value: 4.592e-16
Рисунок 0.37 – Диаграмма центра кластера 2 для случая 6 кластеров
[1] "Points: 97"
Call:
lm(formula = overall
durability + eou + battery + lag + picture,
data = x)
Residuals:
Min 1Q Median 3Q Max
-1.58775 -0.35900 -0.01853 0.40862 1.36684
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) -1.76201 0.49361 -3.570 0.000573 ***
durability 0.61095 0.08864 6.893 6.99e-10 ***
eou 0.26510 0.07555 3.509 0.000702 ***
battery 0.08284 0.06548 1.265 0.209085
lag 0.10125 0.08009 1.264 0.209390
picture 0.36997 0.07894 4.687 9.69e-06 ***
---
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
Residual standard error: 0.605 on 91 degrees of freedom
(46 observations deleted due to missingness)
Multiple R-squared: 0.6329, Adjusted R-squared: 0.6127
F-statistic: 31.38 on 5 and 91 DF, p-value: < 2.2e-16
Рисунок 0.38 – Диаграмма центра кластера 3 для случая 6 кластеров
[1] "Points: 27"
Call:
lm(formula = overall
durability + eou + battery + lag + picture,
data = x)
Residuals:
Min 1Q Median 3Q Max
-0.74858 -0.42160 0.06625 0.40068 0.74995
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) -1.91034 0.63386 -3.014 0.00661 **
durability 0.45722 0.15146 3.019 0.00653 **
eou 0.34519 0.13036 2.648 0.01505 *
battery 0.09205 0.11323 0.813 0.42537
lag 0.28180 0.11826 2.383 0.02670 *
picture 0.33937 0.14131 2.402 0.02566 *
---
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
Residual standard error: 0.5273 on 21 degrees of freedom
(13 observations deleted due to missingness)
Multiple R-squared: 0.8275, Adjusted R-squared: 0.7865
F-statistic: 20.15 on 5 and 21 DF, p-value: 2.264e-07
Рисунок 0.39 – Диаграмма центра кластера 4 для случая 6 кластеров
[1] "Points: 302"
Call:
lm(formula = overall
durability + eou + battery + lag + picture,
data = x)
Residuals:
Min 1Q Median 3Q Max
-3.2823 -0.3548 0.0185 0.4449 1.5182
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) -1.08922 0.24888 -4.377 1.67e-05 ***
durability 0.37473 0.05443 6.884 3.48e-11 ***
eou 0.21288 0.04923 4.324 2.09e-05 ***
battery 0.07361 0.04016 1.833 0.0678 .
lag 0.21206 0.05146 4.121 4.90e-05 ***
picture 0.43024 0.05296 8.124 1.23e-14 ***
---
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
Residual standard error: 0.6593 on 296 degrees of freedom
(69 observations deleted due to missingness)
Multiple R-squared: 0.6346, Adjusted R-squared: 0.6284
F-statistic: 102.8 on 5 and 296 DF, p-value: < 2.2e-16
Рисунок 0.40 – Диаграмма центра кластера 5 для случая 6 кластеров
[1] "Points: 233"
Call:
lm(formula = overall
durability + eou + battery + lag + picture,
data = x)
Residuals:
Min 1Q Median 3Q Max
-1.81123 -0.35567 0.00992 0.38592 1.53587
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) -0.79927 0.21685 -3.686 0.000285 ***
durability 0.39870 0.05865 6.798 9.25e-11 ***
eou 0.20772 0.04877 4.259 3.01e-05 ***
battery 0.12897 0.03901 3.306 0.001099 **
lag 0.11806 0.04680 2.523 0.012335 *
picture 0.40996 0.05130 7.991 6.72e-14 ***
---
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
Residual standard error: 0.5655 on 227 degrees of freedom
(84 observations deleted due to missingness)
Multiple R-squared: 0.7415, Adjusted R-squared: 0.7358
F-statistic: 130.2 on 5 and 227 DF, p-value: < 2.2e-16
Рисунок 0.41 – Диаграмма центра кластера 6 для случая 6 кластеров
[1] "Points: 497"
Call:
lm(formula = overall
durability + eou + battery + lag + picture,
data = x)
Residuals:
Min 1Q Median 3Q Max
-2.94389 -0.49390 0.02783 0.49587 1.92728
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) -0.75873 0.25869 -2.933 0.00351 **
durability 0.39434 0.04881 8.079 5.11e-15 ***
eou 0.13192 0.04024 3.278 0.00112 **
battery 0.15877 0.03600 4.410 1.27e-05 ***
lag 0.13122 0.04339 3.024 0.00263 **
picture 0.37319 0.04128 9.041 < 2e-16 ***
---
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
Residual standard error: 0.6969 on 491 degrees of freedom
(206 observations deleted due to missingness)
Multiple R-squared: 0.4808, Adjusted R-squared: 0.4755
F-statistic: 90.93 on 5 and 491 DF, p-value: < 2.2e-16
Сравнение групповых предпочтений
Рисунок 0.42 – График предпочтений по надежности для 6 кластеров
Рисунок 0.43 – График предпочтений по автономной работе для 6 кластеров
Рисунок 0.44 – График предпочтений по скорости срабатывания для 6 кластеров
Рисунок 0.45 – График предпочтений по простоте использования для 6 кластеров
Рисунок 0.46 – График предпочтений по качеству снимка для 6 кластеров
|
|
|
Скачать 2.87 Mb.