To accurately identify the temperature compensation model of humidity sensitive capacitor

and reduce error of system measurement

the principle of temperature compensation of humidity sensitive capacitor is analyzed.A kind of improved compensation algorithm combined genetic algorithm（GA） with support vector machine（SVM） is proposed.The GA-SVM compensation model is established based on the humidity sensitive capacitor

and compared with BP neural network method.A total of 144 sample point

after compensation

the number relative error absolute value of GA-SVM and BP model less than 3 % is 143

110.The maximum value is 0.34 %

19.68 %.The results show that the improved method can compensate temperature influence effectively

the measurement precision of humidity sensitive capacitor is improved.At the same time

the approximation ability and generalization ability of this algorithm is superior to BP neural network.Therefore

using this method for temperature compensation of humidity sensitive capacitor sensor is effective and feasible.