calculate value based on another value

here is the dataset

Catogory,Confidence,analysis,Level
business, 0.848333, positive, 0.0706335
business, 0.510236, positive, 0.0102158
business, 0.852006, negative, -0.0314569
business, 0.834497, negative, -0.0182176
business, 0.835959, negative, -0.0188398
business, 0.848469, negative, -0.0120408
business, 0.897209, negative, -0.0495422

Confidence give you the surety that analysis is .84 or .51 true. On the bases of confidence I want to calculate result.

I just want my result in such a way that

Business:
positive: some value(confidence) Level:some value
negative: some value(confidence) Level: some value

chamach said:

here is the dataset

Catogory,Confidence,analysis,Level
business, 0.848333, positive, 0.0706335
business, 0.510236, positive, 0.0102158
business, 0.852006, negative, -0.0314569
business, 0.834497, negative, -0.0182176
business, 0.835959, negative, -0.0188398
business, 0.848469, negative, -0.0120408
business, 0.897209, negative, -0.0495422

Confidence give you the surety that analysis is .84 or .51 true. On the bases of confidence I want to calculate result.

I just want my result in such a way that

Business:
positive: some value(confidence) Level:some value
negative: some value(confidence) Level: some value

Click to expand...

The question is not at all clear to me - so my suggestions may be way off the mark. I think you may be trying to interpolate, or extrapolate, or make a model to fit the data.

It appears that "positive" and "negative" are two separate models.

For the positive case, you only have two data points, so the only thing you can do is make a straight line. No way to estimate the statistical probability that the straight line is accurate.

For the negative case, the data don't immediately show a correlation - lets look at the relative ranks of confidence and of level:
business, 0.852006, rank 4, negative, -0.0314569, rank 4
business, 0.834497, rank 1, negative, -0.0182176, rank 2
business, 0.835959, rank 2, negative, -0.0188398, rank 3
business, 0.848469, rank 3, negative, -0.0120408, rank 1
business, 0.897209, rank 5. negative, -0.0495422, rank 5

Well, that does look hopeful that a correlation exists. You should make a plot of the data points, and see if it makes sense to draw a straight line that is an estimate of the data. You could also do a linear regression to estimate such a line. Is that what you need?