Re: Data requirements for factor analysis

Posted by Art Kendall on Jul 03, 2012; 12:46pm
URL: http://spssx-discussion.165.s1.nabble.com/Data-requirements-for-factor-analysis-tp5713939p5713977.html

I still don't have a clear idea of what your data looks like.
But it does not sound anything like a factor analysis so far.
Different disciplines use the term "factor" differently.  It sounds like what you call a factor would be a predictor or independent variable.
Business would be called a criterion, predicted, left hand side, or dependent variable. It might be in currency (dollars, pounds, euros, etc.) number of sales, etc.
Accidents would be called a predictor, independent variable, right hand side, or covariate. Do you have any variables about the severity of the accidents?
Given that you have dates, it sounds like some form of repeated measures MAYBE with enough times to use a time series approach

Factor analysis deals with reducing many variables in a set to a fewer number of dimensions that explain a lot of the variance within the set.
Predictive models have at least 2 sets of variables. One set are those that are predicted or explained. And another set of variables are used to predict or explain the first set. What kind of predictive model is used depends on the number of variables in each set and their levels of measurement, etc.

What entity does a record represent?
What other columns do you have?
How many records do you have all together?

On 7/2/2012 7:18 PM, mawais31 wrote:

Dear Art Kendall!

You are right, the actuall problem is that there is decrease in business due
to decrease in accidents. and there are 6000 records /month...

How many time points do you have?

*I have only date column with following format yyyymmdd. *

Are you then saying that mileage, weather, insurance company, etc influence
the amount of business which in turn explains the number of accidents?

*the business is for body and paint parts etc. so if there is decrease of
accidents then there would be decrease in business *

I haven't weather data but I will put it from another database... The
problem is that whether I use factor analysis or some other technique? or
what procedure I follow to isolate factors... as there might be more factors
such as city safety?




--
View this message in context: http://spssx-discussion.1045642.n5.nabble.com/Data-requirements-for-factor-analysis-tp5713939p5713966.html
Sent from the SPSSX Discussion mailing list archive at Nabble.com.

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD

===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD