Power transaction process are competitive in restructured markets. Market operations such as bidding, hedging, planning of facility investments and negotiation of bilateral contracts are rely on electricity price. Electricity price uncertainty is the major barrier in profit maximization of market participants. This price uncertainty can be addressed by using accurate forecasts in market operations. Artificial intelligence and time series-based models are used for day ahead electricity price forecasting. Such models necessitate massive training inputs for accurate forecasting, which leads to high processing time. However, fast and accurate forecasts are required for real time decision making in market operations. Therefore, this paper proposes three support vector regression forecasting models for day ahead electricity price forecasting. Proposed models use limited quantum of training inputs as they use similar day approach for forecasting. Also, proposed model use GA and PSO based hyper parameter optimization for improving forecasting accuracy. Data collected from Indian power market is used for performance evaluation. Performance of proposed models are compared with linear regression and artificial neural network models. Results show that proposed models have very strong potential towards day ahead electricity price forecasting.