why do we need a moving window integrator in pan-tompkins algorithm , and what is meant by width -information of qrs complex , and how does this latter help us in QRS detection ?
Without the moving average filter, the Q and S parts would not contribute at all to the detection of the QRS complex, because they wouldn't affect the size or position of the R peak. With the moving average filter, the R peak is made higher by the Q and S parts around it.
The "width" of the QRS complex is its duration, i.e., its width on a ECG graph. The length of the moving average filter needs to be about this long to best accomplish the above purpose without adding P or T segments to the peak.