Dear researcher, baseline stability means that the recorded signal for each addition begins from almost the same baseline signal, response time has to deal with the fact how fast your sensor reacts in the addition of your analyte (or in changes of the analyte's concentration) and finally recovery time means how fast the sensor goes back to the baseline signal (small relaxation time) in order to perform the next measurement.

Dear researcher, you can try with the classical Newton's Cooling Law. It very useful for temperature sensors.

Thank you very much for your comments. The baseline is the sharp spike in the picture that I attach. It just happens that the baseline is stable and straight. In some cases, the baseline is not stable like Pt/WO3 sensor. The response time is the time taken to achieve 90% of the equilibrium point when gas/analyte is introduced - which calculated from 90% change in resistance, then spot the time when the resistance has changed by 90%. Similarly, the recovery time is calculated that way. 

As far as I have seen in publications the response and recovery times are analogous to rise and fall times in capacitors, which use the difference between 10% - 90% of signal change. The baseline is as mentioned above the time before the signal response where steady state is said to be achieved.  Baselines tend to drift overtime for chemical sensors due to change in the sensing material (e.g. oxidation).  In the attached plot the responses are very sharp and don't appear to have reached saturation.  The response and recovery can provide information about the chemical that is interacting with the sensing material (e.g. dipole moment, molecular weight)

Given a sensor you would want to expose it to various concentrations to the chemical analtye until it saturates the response and produces a flat response.  The peaks of each of these concentration dependent curves can be use with the Stern-Volmer relationship to predict low-level detection limits.