When writing patents, look on the use of discrete data structures used to predict molecular interactions in literature, the sensitivity and specificity also a bit concerned about.
Prior to cooperate when papers and brothers, just rote TP, TN and other indicators, but carefully recall when the brothers actually been previously discussed and laboratory students with examples illustrate this problem, and an international conference on computing services have been other teacher asks a question speakers this problem.
Sensitivity sensitivity, that is true positive rate, indicating that the proportion of really sick of all the samples in the sample forecasting sick.
For specific specificity, it is a true negative rate, indicating that the disease is not real proportion of the sample did not predict a percentage of all disease samples.
Due to the sensitivity if the forecast is sick but not sick, not as to minimize the importance of this disease but did not predict the actual diseased sample rate. This shows that, in actual diagnosis expert system, the two indicators is more important is the specificity.
sensitivity = recall = true positive rate
specificity = 1- false positive rate