如何在Ruby中找到标准偏差？

Question

我写了一个程序,它在一个单独的txt文件中查找数据,然后给出平均值和标准偏差.它找到了我的平均值,但我得到标准偏差的错误.想知道有谁可以帮我修复我的代码.就是这个:

data = File.open("avg_temp.txt", "r+")

contents = data.read

contents = contents.split("\r\n")
#split up array
contents.collect! do |x|
    x.split(',') 
end


sum = 0

contents.each do |x|
#make loop to find average
        sum = sum  + x[1].to_f
end
avg = sum / contents.length
puts "The average temperature of Laguardia Airport from 11/97 - 05/11 is:
#{ avg.round(3)}C (Answer is rounded to nearest thousandth place)"
#puts average


variance = 0
contents.each do |x|
        variance = variance + (x-avg)**2
end

variance = variance / contents
variance = Math.sqrt(variance)
puts variance

我在第27行遇到错误:variance = variance +(x-avg)**2

avg_temp.rb:27:in `-': can't convert Float into Array (TypeError)
    from avg_temp.rb:27:in `block in <main>'
    from avg_temp.rb:26:in `each'
    from avg_temp.rb:26:in `<main>'

Answer 1

正如您的问题已得到解答,也许我可以建议一种使用Ruby来计算样本均值和标准差的典型方法:

contents = [1,2,3,4,5,6,7,8,9]
n = contents.size             # => 9
contents.map!(&:to_f)         # => [1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0]
mean = contents.reduce(&:+)/n # => 5.0
sum_sqr = contents.map {|x| x * x}.reduce(&:+) # => 285.0
std_dev = Math.sqrt((sum_sqr - n * mean * mean)/(n-1)) # => 2.7386127875258306

• 'contents'包含从文件中获取的样本值
• 我们也可以将所有样本值转换为浮点数,就像你所做的那样; 否则我们将不得不做类似于=(x1 + x2 + ...)/ n.to_f的事情,这有点混乱,我们可能忘记在任何地方都这样做.contents.map(&:to_f)与contents.map {| x |相同 x.to_f}.('collect'是'map'的另一个名字.)
• contents.reduce(&:+)与contents.reduce {| tot,x |相同 tot + x},所以它只是将值相加.('inject'是'reduce'的另一个名字.)我们用n除以得到均值.
• contents.map {| x | x*x}生成一个数组[1,4,9,16,...,81].
• 然后我们将其链接到[1,4,9,16,...,81] .reduce(&:+),它将平方值相加.
• 在计算样本标准偏差时,除以n-1.(除以n来计算总体标准偏差.)如果你想将其四舍五入到比如说2个小数,则坚持.round(2)在最后一个语句的末尾(紧跟在最后一个权利之后).