Method for estimating RSS missing value based on adaptive context generative adversarial model

The invention discloses a method for estimating an RSS missing value based on an adaptive context generative adversarial model. The method comprises the following steps of (1) establishing an adaptivecontext generative adversarial model, namely, an ACOGAN model, in combination with an Auto Encoder model and a GAN model, wherein the ACOGAN model comprises a generator and a discriminator, and the generator is formed by connecting an encoder and a decoder through a channel full connection layer; (2) generating RSS fingerprint simulation data through the ray tracing technology to serve as a training set and a test set of an ACOGAN model; (3) preprocessing the training set, and converting the training set into an input format required by an ACOGAN model; (4) training an ACOGAN model; (5) exporting training parameters of the ACOGAN model; (6) preprocessing the test set, and converting the test set into an input format required by the ACOGAN model; and (7) predicting the RSS fingerprint of the specific position with