| 摘要 |
A major shortcoming in the prior art for detecting cardiac arrhythmias for the purpose of making an electrical intervention decision has been that each variable sensed was given in effect a "veto" over intervention on the basis of its own particular predetermined threshold value, and irrespective of the values of companion variables. The present invention overcomes this prior art problem by generating a function of relevant variables, and assigning a threshold value to the function. As a result, an individual variable can range widely for a positive intervention decision, depending upon the values of the companion variables. Thus, the intervention exploits features often described as those associated with fuzzy logic. The functions chosen for this purpose can be as one desires, with the choice depending upon the current state of advancing knowledge. Even further flexibility can be achieved by employing the principles of the neural net, wherein functional values can themselves be combined, with another threshold value assigned to the result of that combination, and so on.
|
