帮忙翻译,送电子软件

这是NOKIA的一个计数型GRR软件的说明,有高手翻译成中文的话,我将软件共享给大家.
Introduction
This worksheet has been designed to collect and analyse data of an Attribute Gauge Repeatability and Reproducibility study.
Attribute data is pass/fail data which would normally be collected, for example, during a visual quality inspection process.
The sheets are of this spreadsheet protected hadorwould password, you can only modify cells that require user input.
The principle of Attribute Gauge R&R is that multiple inspectors (human or machine) categorise a set of samples into pass or
fail according to their own inspection criteria. This inspection process is carried out twice on the same set of samples in
order to verify the Repeatability of the inspection process. Internal consistency of the inspectors is analysed together with
the external consistency against a known inspection decision decided on by experts.


This workbook enables you to take up to 12 operators into consideration at the same time, and up to 150 part samples.
Separate sheets give detailed information about samples and operators.

Sheet "Collect Data"
If the operators need to report their inspection decision on a piece of paper, the 'Collect Data' sheet can be printed for each
of the trials and operators. In this way the operators can be prevented from knowing the expert decision (Attribute) or the
results of their previous trial. This helps to reduce bias.
The reasons for rejection of any particular sample can also be collected on this sheet by the operators. From this
information the sample set can be revised or training plans for the operators devised.

Sheet "Master Data"
Enter the expert's decision (Attribute) in this table. The description of the pass / fail condition also needs to be entered, e.g.
good / bad or accept / reject, etc. Only those names will be accepted in cells of the other sheets.

Sheet "Data Entry"
Once the operators' first and second inspection results have been collected, they need to be entered into the 'Data Entry'
sheet. You should give the names of operators in the cell after that called: 'Name'.
All calculations are done automatically, and you will find results in following sheets:

Sheet "Summary"
The results of each operator are automatically copied from the 'Data Entry' sheet and then compared to the expert's
Decision from the 'Master Data' sheet. The first column headed 'Sample No.' gives the reference number for each sample part.
In the second column, the Expert's Decision for each part is shown. In the third column it can be seen cumulatively if all operators
agree or disagree with the expert decision. In the fourth, fifth and sixth columns, the cumulative number of 'False Negative'
(when an operator rejects a truly good part on both the first and second trials), 'Mixed' (when an operator has a different
inspection decision on the first and second trials) and 'False Positive' (when an operator passes a truly bad part on both the
first and second trials) are shown.
Subsequent columns give the results of each operator and then show a 'Summary' of each result for a particular sample
number. The summary can give one of the following descriptions: 'Agree', 'Mixed/Fail', 'Mixed/Pass', 'False -ve' or
False +ve'. Rows at bottom of the sheet show the cumulative summary across all operators for all parts, and the cumulative
summary of all data for a particular operator.

Sheet "Roll-up"
This sheet shows the number of samples inspected by each operator, the number of agrees with the experts, the number
of consistent decisions (same first and second trial) that did not match the expert's decision, (i.e. false negatives and false
positives) and the number of mixed inspections (the number of times the operator changed his/her mind on trial one and
two). In normal inspection, operators do not inspect each part twice, so it can be assumed that they will reject 50% and
accept 50% of those parts with mixed decisions. This is a approximation of real conditions, but allows the mixed decisions
to be accounted for in the analysis. The number of false negatives including the effect of the mixed inspections is the sum
of the number of false negatives and half of the number of mixed answers with a truly good part.
The number of false positives including the effect of the mixed inspections is the sum of the number of false positives and
half of the number of mixed answers with a truly bad part.
The total percentage of rejected good parts and accepted bad parts are calculated on the basis of these figures.

The next three columns shows '%Appraiser' which is the internal efficiency of the operators on the first and second trials.
The '%Appraiser' result gives no information about how the operators performed in relation to the expert's decision, only
how consistent they were over the two trials. Remember, they could have been consistently wrong! Also shown are the
95% confidence intervals for the '%Appraiser' results. These indicate the upper and lower percentages that a particular
operator could be operating at with a 95% statistical confidence. Pay particular attention to the lower 95% confidence interval.

The next three columns shows '%Score vs Attribute' which is how reliable the operators were on the first and second trials
and how well they matched the experts' decision. For example, a result of 100% indicates that a particular operator had the
same decision on both the first and second trial, and all decisions for all parts agreed with the experts' decisions.
Thus the '%Score vs Attribute' represents the external efficiency of each operator.
95% Confidence intervals for '%Score vs Attribute' are also given. Again the 95% lower confidence interval should be
carefully checked as this indicates how bad a particular operator performance could be.
The bottom row of the table shows the 'Screen' values. These values show the effect of all operators combined.

For example:

Screen Sample = Total number of samples in study
Screen Agree = Total number of samples where all operators agreed on both trials
Screen False Negative = Total number of samples which all operators agreed were false negatives
Screen False Positive = Total number of samples which all operators agreed were false negatives
Screen Mixed when Fail/Pass = Total number of samples which all operators agreed a mixed inspection
Screen False Negative (Incl. Mixed) = Total number of false negatives including the effect of the mixed inspections
Screen False Positive (Incl. Mixed) = Total number of false positives including the effect of the mixed inspections
Screen %Appraiser = The internal efficiency of all operators across both trials (note also Screen values for the 95% Confidence
Intervals
Screen %Score vs. Attribute = The external efficiency of all operators against the expert decision across both trials (note also
the Screen values for the 95% Confidence Intervals)

Sheet "Graphs"
There are 4 graphs presented on the graph sheet:
- %Appraiser
- %Score vs Appraiser
- False Answers by Operator
- False Answers by Sample Number

The %Appraiser graph shows the %Appraiser results for each operator, i.e. the internal efficiency of each operator on both trials but
not against the expert's decision. Also displayed is the Screen %Appraiser result. The actual score is represented by the green
diamond symbol on the graph. The 95% Confidence Intervals are represented by red horizontal lines. The upper and lower 95%
confidence intervals are joined by a vertical line indicating the range of possible scores with a 95% statistical confidence.
Any crossing of a particular operator's score with the 95% Confidence Intervals of another operator indicate no statistical difference.

The %Score vs Attribute graph shows the %Score vs Attribute results for each operator, i.e. the external efficiency of each operator
on both trials against the expert's decision. Also displayed is the Screen %Score vs Attribute result. The actual score is represented
by the green diamond symbol on the graph. The 95% Confidence Intervals are represented by red horizontal lines. The upper and
lower 95% confidence intervals are joined by a vertical line indicating the range of possible scores with a 95% statistical confidence.
Any crossing of a particular operator's score with the 95% Confidence Intervals of another operator indicate no statistical difference.

The False Answers by Operator graph shows the frequency of the different types of false answers given by each operator. The following
types of false answers displayed are:
- False Negative (Operator rejects a truly good part on both trials)
- False Positive (Operator accepts a truly bad parts on both trials)
- Mixed When Fail (Operator has a mixed inspection result on a truly bad part)
- Mixed When Pass (Operator has a mixed inspection result on a truly good part)

The False Answers by Sample Number graph shows the frequency of the different types of false answers given for each part. The types
of false answer displayed are:
- False Negative (Operator rejects a truly good part on both trials)
- False Positive (Operator accepts a truly bad parts on both trials)
- Mixed Answers (Operator gives a different inspection decision on each trial)