thongkemota
thongke.club
7/30/2020
Kết nối dữ liệu
options(scipen=999)
library(readxl)
setwd("d:/DATA2020/HuongGian_FindModel")
dulieu <-read_excel("data.editor.SME.xlsx")
head(dulieu)## # A tibble: 6 x 36
## Year Firm NHOM NS LnNS Labor LnLab IMS LnIMS ShortASS_old
## <dbl> <chr> <dbl> <dbl> <dbl> <dbl> <dbl> <chr> <dbl> <dbl>
## 1 2015 Kiml~ 1 14343 9.57 70 4.25 85 4.44 14603
## 2 2016 Kiml~ 1 35749 10.5 70 4.25 85.4 4.45 25637
## 3 2017 Kiml~ 1 57619 11.0 70 4.25 89 4.49 40326
## 4 2018 Kiml~ 1 58094 11.0 70 4.25 81.8 4.40 59317
## 5 2019 Kiml~ 1 82752 11.3 70 4.25 87 4.47 61195
## 6 2015 Wint~ 2 13441. 9.51 35 3.56 82.5 4.41 23974.
## # ... with 26 more variables: LongASS_old <dbl>, ShortLIA_old <dbl>,
## # LongLIA_old <dbl>, Equity_old <dbl>, ShortASS <dbl>, LongASS <dbl>,
## # ShortLIA <dbl>, LongLIA <dbl>, Equity <dbl>, LnSASS <dbl>, LnLASS <dbl>,
## # LnSLIA <dbl>, LnLLIA <dbl>, LnEQU <dbl>, ICTinf <dbl>, ICThum <dbl>,
## # ICTapp <dbl>, LnINF <dbl>, LnHUM <dbl>, LnAPP <dbl>, TRI <dbl>, ESS <dbl>,
## # GGDP <dbl>, LnESS <dbl>, LnGDP <dbl>, Dummy <dbl>Tách dữ liệu cần thiết
library(tidyverse)## -- Attaching packages ------------------------------------------------------------ tidyverse 1.3.0 --## v ggplot2 3.3.2 v purrr 0.3.4
## v tibble 3.0.1 v dplyr 1.0.0
## v tidyr 1.1.0 v stringr 1.4.0
## v readr 1.3.1 v forcats 0.5.0## -- Conflicts --------------------------------------------------------------- tidyverse_conflicts() --
## x dplyr::filter() masks stats::filter()
## x dplyr::lag() masks stats::lag()dulieu2 <-dulieu %>% select(LnNS, LnLab , LnIMS , ShortASS , LongASS , ShortLIA , LongLIA , Equity , ICTinf , ICThum , ICTapp ,TRI)Thống kê mô tả
summary(dulieu2)## LnNS LnLab LnIMS ShortASS
## Min. : 7.093 Min. :0.6931 Min. :4.248 Min. : 0.1452
## 1st Qu.:10.082 1st Qu.:2.7081 1st Qu.:4.458 1st Qu.: 1.4509
## Median :11.093 Median :3.5553 Median :4.496 Median : 3.2568
## Mean :11.186 Mean :3.7285 Mean :4.500 Mean : 11.4877
## 3rd Qu.:12.230 3rd Qu.:4.5951 3rd Qu.:4.543 3rd Qu.: 7.9543
## Max. :15.505 Max. :9.8995 Max. :4.605 Max. :271.6951
## NA's :2 NA's :2
## LongASS ShortLIA LongLIA Equity
## Min. : 0.000 Min. : 0.0045 Min. : 0.000 Min. :-0.5390
## 1st Qu.: 0.103 1st Qu.: 0.8426 1st Qu.: 0.000 1st Qu.: 0.5335
## Median : 0.611 Median : 2.4667 Median : 0.000 Median : 1.3278
## Mean : 3.706 Mean : 9.1068 Mean : 1.247 Mean : 4.8441
## 3rd Qu.: 3.442 3rd Qu.: 6.1322 3rd Qu.: 0.455 3rd Qu.: 3.8166
## Max. :52.431 Max. :217.9357 Max. :30.031 Max. :83.4522
##
## ICTinf ICThum ICTapp TRI
## Min. :0.4536 Min. :0.6804 Min. :0.3950 Min. :5.830
## 1st Qu.:0.5766 1st Qu.:0.7246 1st Qu.:0.5932 1st Qu.:6.040
## Median :0.5892 Median :0.7452 Median :0.6561 Median :6.140
## Mean :0.5733 Mean :0.7400 Mean :0.6087 Mean :6.184
## 3rd Qu.:0.6052 3rd Qu.:0.7720 3rd Qu.:0.6574 3rd Qu.:6.310
## Max. :0.6419 Max. :0.7778 Max. :0.7420 Max. :6.600
## library(table1)##
## Attaching package: 'table1'## The following objects are masked from 'package:base':
##
## units, units<-table1(~LnNS + LnLab + LnIMS + ShortASS + LongASS + ShortLIA + LongLIA + Equity + ICTinf + ICThum + ICTapp + TRI, dulieu2)| Overall (N=825) |
|
|---|---|
| LnNS | |
| Mean (SD) | 11.2 (1.57) |
| Median [Min, Max] | 11.1 [7.09, 15.5] |
| LnLab | |
| Mean (SD) | 3.73 (1.34) |
| Median [Min, Max] | 3.56 [0.693, 9.90] |
| Missing | 2 (0.2%) |
| LnIMS | |
| Mean (SD) | 4.50 (0.0623) |
| Median [Min, Max] | 4.50 [4.25, 4.61] |
| Missing | 2 (0.2%) |
| ShortASS | |
| Mean (SD) | 11.5 (28.4) |
| Median [Min, Max] | 3.26 [0.145, 272] |
| LongASS | |
| Mean (SD) | 3.71 (7.12) |
| Median [Min, Max] | 0.611 [0, 52.4] |
| ShortLIA | |
| Mean (SD) | 9.11 (23.1) |
| Median [Min, Max] | 2.47 [0.00450, 218] |
| LongLIA | |
| Mean (SD) | 1.25 (3.72) |
| Median [Min, Max] | 0 [0, 30.0] |
| Equity | |
| Mean (SD) | 4.84 (10.1) |
| Median [Min, Max] | 1.33 [-0.539, 83.5] |
| ICTinf | |
| Mean (SD) | 0.573 (0.0638) |
| Median [Min, Max] | 0.589 [0.454, 0.642] |
| ICThum | |
| Mean (SD) | 0.740 (0.0354) |
| Median [Min, Max] | 0.745 [0.680, 0.778] |
| ICTapp | |
| Mean (SD) | 0.609 (0.117) |
| Median [Min, Max] | 0.656 [0.395, 0.742] |
| TRI | |
| Mean (SD) | 6.18 (0.260) |
| Median [Min, Max] | 6.14 [5.83, 6.60] |
Cách tính độ nhọn độ nghiêng
# Trong thống kê mô tả không nhất thiết phải đưa 2 chỉ số này, nhưng em yêu cầu thì anh làm luôn
library(moments)
skewness(dulieu2$LnNS)## [1] 0.2589435skewness(dulieu2$LnLab)## [1] NAskewness(dulieu2$LnIMS)## [1] NAskewness(dulieu2$ShortASS)## [1] 5.333304skewness(dulieu2$LongASS)## [1] 3.021975skewness(dulieu2$ShortLIA)## [1] 5.163209skewness(dulieu2$LongLIA)## [1] 4.568951skewness(dulieu2$Equity)## [1] 4.290422skewness(dulieu2$ICTinf)## [1] -1.046936skewness(dulieu2$ICThum)## [1] -0.578595skewness(dulieu2$ICTapp)## [1] -0.8996106skewness(dulieu2$TRI)## [1] 0.3036691kurtosis(dulieu2$LnNS)## [1] 2.79321kurtosis(dulieu2$LnLab)## [1] NAkurtosis(dulieu2$LnIMS)## [1] NAkurtosis(dulieu2$ShortASS)## [1] 36.227kurtosis(dulieu2$LongASS)## [1] 13.3092kurtosis(dulieu2$ShortLIA)## [1] 33.79418kurtosis(dulieu2$LongLIA)## [1] 26.54634kurtosis(dulieu2$Equity)## [1] 25.17663kurtosis(dulieu2$ICTinf)## [1] 2.768744kurtosis(dulieu2$ICThum)## [1] 2.005576kurtosis(dulieu2$ICTapp)## [1] 2.586861kurtosis(dulieu2$TRI)## [1] 2.041062Vẽ đồ thị boxplot ( Nếu cần)
boxplot(dulieu2$LnNS)
boxplot(dulieu2$LnLab)
boxplot(dulieu2$LnIMS)
boxplot(dulieu2$ShortASS)
boxplot(dulieu2$LongASS)
boxplot(dulieu2$ShortLIA)
boxplot(dulieu2$LongLIA)
boxplot(dulieu2$Equity)
boxplot(dulieu2$ICTinf)
boxplot(dulieu2$ICThum)
boxplot(dulieu2$ICTapp)
boxplot(dulieu2$TRI)
Vẽ đồ thị Histogram ( Nếu cần)
hist(dulieu2$LnNS)
hist(dulieu2$LnLab)
hist(dulieu2$LnIMS)
hist(dulieu2$ShortASS)
hist(dulieu2$LongASS)
hist(dulieu2$ShortLIA)
hist(dulieu2$LongLIA)
hist(dulieu2$Equity)
hist(dulieu2$ICTinf)
hist(dulieu2$ICThum)
hist(dulieu2$ICTapp)
hist(dulieu2$TRI)