我想为多个数据框创建一个具有ifelse()条件的列。在这种情况下,数据帧是用于加密货币的3个时间序列数据。这是自动下载3个数据框的代码:
INotifyPropertyChanged
现在,我想在返回中检测到积极的过度反应(oR_pos)。我将过度反应定义为高于平均值+ 1标准偏差的值(返回值)。我也想针对1.5和2个标准差进行此操作。这是我想要的一种加密货币(比特币)的输出:
library(tidyverse)
library(crypto)
crypto_chart <- crypto_prices()%>% select(-id, -symbol,-price_btc, -`24h_volume_usd`,-available_supply, -total_supply,-max_supply, -percent_change_1h, -percent_change_24h, -percent_change_7d, -last_updated)%>% slice(1:3)
list_cryptocurrencies <-crypto_chart$name
map(list_cryptocurrencies,
function(x) crypto_history(x, start_date = '20150101', end_date = '20190303')%>%
select(-slug, -symbol, -name, -`ranknow`))%>%
set_names(list_cryptocurrencies)%>%
list2env(envir = .GlobalEnv)
##Calculating return
map(mget(list_cryptocurrencies),
function(x) x %>% mutate(`return` = (close-open)/open * 100))%>%
list2env(mget(list_cryptocurrencies), envir = .GlobalEnv)
现在我有3个新列,它们的反应过度(oR_pos)> 1sd; 1.5sd和2sd。
我已经尝试过此代码:
> Bitcoin
date open close return oR_pos>1sd oR_pos>1.5sd oR_pos>2sd
1 2018-01-01 14112.2 13657.2 -3.2241607 NA NA NA
2 2018-01-02 13625.0 14982.1 9.9603670 9.960367 9.960367 9.960367
3 2018-01-03 14978.2 15201.0 1.4874952 NA NA NA
4 2018-01-04 15270.7 15599.2 2.1511784 NA NA NA
5 2018-01-05 15477.2 17429.5 12.6140387 12.614039 12.614039 12.614039
6 2018-01-06 17462.1 17527.0 0.3716621 NA NA NA
7 2018-01-07 17527.3 16477.6 -5.9889430 NA NA NA
8 2018-01-08 16476.2 15170.1 -7.9271919 NA NA NA
9 2018-01-09 15123.7 14595.4 -3.4931928 NA NA NA
10 2018-01-10 14588.5 14973.3 2.6376941 NA NA NA
11 2018-01-11 14968.2 13405.8 -10.4381288 NA NA NA
12 2018-01-12 13453.9 13980.6 3.9148500 3.914850 NA NA
但是它不起作用。 有人可以帮我吗?
答案 0 :(得分:1)
以下内容与您的预期功能紧密匹配,将所需的列添加到您的加密货币上,同时允许将所需的sd阈值作为参数传入以提高灵活性。顺便提一句,下面的解决方案根据OP使用>
,但是您可能希望考虑从sd的移动+/-方向。可以改为使用以下解决方案:
col <- ifelse(returns > (r_mean+(r_sd*threshold)) |
returns < (r_mean-(r_sd*threshold)),
returns,NA)
解决方案如下:
oR_pos_function <- function(returns,thresholds) {
r_mean <- mean(returns,na.rm=T)
r_sd <- sd(returns,na.rm=T)
cols <- lapply(thresholds,function(threshold) {
col <- ifelse(returns > (r_mean+(r_sd*threshold)),returns,NA)
return(col)
})
cols <- as.data.frame(cols)
names(cols) <- paste0("oR_pos>",thresholds,"sd")
return(cols)
}
new_cols <- oR_pos_function(returns=Bitcoin$return,thresholds=c(1,1.5,2))
Bitcoin <- cbind(Bitcoin,new_cols)
结果:
> head(Bitcoin[Bitcoin$date>="2018-01-01",])
date open high low close volume market close_ratio spread return oR_pos>1sd oR_pos>1.5sd oR_pos>2sd
1097 2018-01-01 14112.2 14112.2 13154.7 13657.2 10291200000 229119155396 0.5248042 957.5 -3.2241607 NA NA NA
1098 2018-01-02 13625.0 15444.6 13163.6 14982.1 16846600192 251377913955 0.7972381 2281.0 9.9603670 9.960367 9.960367 9.960367
1099 2018-01-03 14978.2 15572.8 14844.5 15201.0 16871900160 255080562912 0.4894961 728.3 1.4874952 NA NA NA
1100 2018-01-04 15270.7 15739.7 14522.2 15599.2 21783199744 261795321110 0.8845996 1217.5 2.1511784 NA NA NA
1101 2018-01-05 15477.2 17705.2 15202.8 17429.5 23840899072 292544135538 0.8898258 2502.4 12.6140387 12.614039 12.614039 12.614039
1102 2018-01-06 17462.1 17712.4 16764.6 17527.0 18314600448 294217423675 0.8043891 947.8 0.3716621 NA NA NA
>
每个注释的替代项:
oR_pos_function <- function(coin_data,thresholds) {
returns <- coin_data$return
r_mean <- mean(returns,na.rm=T)
r_sd <- sd(returns,na.rm=T)
cols <- lapply(thresholds,function(threshold) {
col <- ifelse(returns > (r_mean+(r_sd*threshold)),returns,NA)
return(col)
})
cols <- as.data.frame(cols)
names(cols) <- paste0("oR_pos>",thresholds,"sd")
coin_data <- cbind(coin_data,cols)
return(coin_data)
}
答案 1 :(得分:0)
您可以使用#include <vector>
添加任何此类字段
dplyr::mutate