Confidence Limits/Confidence Interval/Prediction Interval

[Mathan]

What is the difference between Confidence Limits, Confidence Interval
and Prediction Interval?

[Allabux Jaffer]

See this example: for the variable Ex i have calculated some descriptive statistics:

 

EX
12
15
16
12
14
15
12
Descriptive Statistics
EX
Mean	13.71428571
Standard Error	0.644178536
Median	14
Mode	12
Standard Deviation	1.704336206
Sample Variance	2.904761905
Kurtosis	-2.154743349
Skewness	0.051940802
Range	4
Minimum	12
Maximum	16
Sum	96
Count	7
Confidence Level(95.0%)	1.576248091

here for the sample mean confidence interval is 13.71+or-1.576 at 95% confidence level(i think confidence limits you mean confidence level only. so its simple, confidence level is % of reliable result, c.Interval is the range of the true value will fall.

 

prediction interval??? have to study...... 

[hazux]

prediction interval is simply the range of predicted values.

[Mathan]

Confidence Interval. The confidence intervals for specific statistics (e.g., means, or regression lines) give us a range of values around the statistic where the "true" (population) statistic can be expected to be located.

Confidence Interval vs. Prediction Interval. In regression, it is possible to predict the value of the dependent variable based on given values of the independent variables. When these values are predicted, it is also possible to calculate confidence intervals and/or prediction intervals for the dependent variable.

The confidence interval gives information on the expected value (mean) of the dependent variable. That is, a confidence interval for a predicted value of the dependent variable gives a range of values around which the "true" (population) mean (of the dependent variable for given levels of the independent variables) can be expected to be located (with a given level of certainty, see also Elementary Concepts).

The prediction interval gives information on individual predictions of the dependent variable. That is, a prediction interval for a predicted value of the dependent variable gives us a range of values around which an additional observation of the dependent variable can be expected to be located (with a given level of certainty

Confidence Limits. The same as Confidence Intervals. In Neural Networks, they represent the accept and reject thresholds, used in classification tasks, to determine whether a pattern of outputs corresponds to a particular class or not. These are applied according to the conversion function of the output variable (One-of-N, Two-state, Kohonen, etc).

 

[Hazux A]

IN SIMPLE TERMS, CONFIDENCE INTERVAL IS THE RANGE OF THE EXISTENCE OF
MEAN VALUE. PREDICTION INTERVAL IS THE RANGE OF THE EXISTENCE OF A NEW
OR PREDICTED VALUE.

TOLERANCE INTERVAL IS THE RANGE OF THE EXISTENCE OF A CERTAIN
PERCENTAGE OF THE OBSERVATIONS AT A CERTAIN CONFIDENCE LEVEL. EG. the
range or interval containing 80% of the values at 95% confidence
level.

On May 26, 9:23 am, Mathan <mathan4s...@gmail.com> wrote:
> *Confidence Interval.* The confidence intervals for specific statistics

> (e.g., means, or regression lines) give us a range of values around the
> statistic where the "true" (population) statistic can be expected to be
> located.
>

> *Confidence Interval vs. Prediction Interval.* In regression, it is possible

> to predict the value of the dependent

> variable<http://www.statsoft.com/textbook/statistics-glossary/i.aspx?#Independent
> vs. Dependent Variables> based on given values of the independent variables.
> When these values are predicted, it is also possible to calculate *confidence
> intervals* and/or *prediction intervals* for the dependent variable.
>
> The confidence interval<http://www.statsoft.com/textbook/statistics-glossary/c.aspx?#Confidence
> Interval General> gives information on the expected value
> (mean<http://www.statsoft.com/textbook/statistics-glossary/m.aspx?#mean>)
> of the dependent variable. That is, a *confidence interval* for a predicted

> value of the dependent variable gives a range of values around which the
> "true" (population) mean (of the dependent variable for given levels of the
> independent variables) can be expected to be located (with a given level of
> certainty, see also Elementary

> Concepts<http://statsoft.com/textbook/elementary-concepts-in-statistics/>
> ).
>
> The *prediction interval* gives information on individual predictions of the
> dependent variable. That is, a *prediction interval* for a predicted value

> of the dependent variable gives us a range of values around which an
> additional observation of the dependent variable can be expected to be
> located (with a given level of certainty

>  *Confidence Limits.* The same as Confidence
> Intervals<http://www.statsoft.com/textbook/statistics-glossary/c.aspx?#Confidence
> Interval General>. In Neural Networks, they represent the accept and reject
> thresholds, used in
> classification<http://www.statsoft.com/textbook/statistics-glossary/c.aspx?#Classifi...>tasks,

> to determine whether a pattern of outputs corresponds to a particular
> class or not. These are applied according to the conversion function of the
> output variable

> (*One-of-N*<http://www.statsoft.com/textbook/statistics-glossary/o.aspx?#One-of-N
> Encoding>, *Two-state*<http://www.statsoft.com/textbook/statistics-glossary/t.aspx?#Two-State>,
> Kohonen<http://www.statsoft.com/textbook/statistics-glossary/k.aspx?#Kohonen
> Training>, etc).