Re: Pls advise: "best fit" statistics?


On 25-Jul-2005, joeu2004@xxxxxxxxxxx wrote:

> I looked at the price of homes sold in the county during
> a paricular period. There were 597. I want to determine
> the mean and sd of those 597 prices.
>
> When I graph the data, I see a nearly-symmetrical
> bell-shaped curve with a very long right tail. If I
> compute the mean and sd based on all 597 data, then
> graph a normal curve that fits those statistics, I get a
> curve that is much flatter and broader and shifted more
> to the right than the original distribution, still with
> a long right tail. That is not surprising, considering
> the length of the right tail.
>
> Based on the 1.5*IQR criterion, I identified 51 prices
> (8.5%) as outliers. If I exclude the outliers from the
> computation of the mean and sd, then graph a normal
> curve that fits the newly-computed mean and sd, I get a
> curve that closely fits the remaining 91.5% of the data.
> Thus, the second normal curve is a "best fit" curve for
> the data.

Why don't you use the median price rather than the mean? This is the way
house prices for an area are usually quoted.

--
Phil Sherrod
(phil.sherrod 'at' sandh.com)
http://www.dtreg.com (decision tree modeling)
http://www.nlreg.com (nonlinear regression)
http://www.NewsRover.com (Usenet newsreader)
http://www.LogRover.com (Web statistics analysis)
.

Relevant Pages

  • Re: Gold Prices This Monday Morning
    ... interest in words and word usage. ... Oil and gold prices are actual figures. ... A hyperbola is a kind of a curve and hyperbolic would refer to the ...
    (rec.collecting.coins)
  • Re: Gold Prices This Monday Morning
    ... interest in words and word usage. ... Oil and gold prices are actual figures. ... A hyperbola is a kind of a curve and hyperbolic would refer to the ...
    (rec.collecting.coins)