Generating Synthetic Data to Reduce Prediction Error of Energy Consumption

Renewable and nonrenewable energy sources are widely incorporated for solar wind that produces electricity without increasing carbon dioxide emissions. Energy industries worldwide trying hard to predict future consumption could eliminate over or under contracting resources unnecessary financing. Machine learning techniques predicting the trending solution overcome challenges faced by companies. The basic need machine algorithms be trained accurate prediction requires a considerable amount of data. Another critical factor is balancing data enhanced prediction. Data Augmentation technique used available training. Synthetic generation new which can improve accuracy models. In this paper, we propose model takes time series as input, pre-processes data, then uses multiple augmentation generative adversarial networks generate synthetic when combined with original reduces error. We TGAN-skip-Improved-WGAN-GP tabular modify TGAN skip connections, WGAN-GP defining consistency term, finally use architecture improved training TGAN-skip. various evaluation metrics visual representation compare performance our proposed model. also measured along mean maximum error generated while different variations augmented mode collapse problem handled it converged faster than existing GAN models generation. experiment result shows combining significantly reduce rate increase consumption.