Forecasting electricity prices plays a key role for stakeholders to construct an electricity business strategy. In recent years, electricity power system structures have undergone changes because of the large-scale penetration of renewable energy sources and the change in consumer behavior. These system changes have led to increased uncertainty in the behavior of electricity prices. This behavior is also influenced by various other factors such as past prices, weather, and consumer lifestyle. This research focuses on the day-ahead forecasting of the most recent electricity spot market prices in the Japan Electric Power Exchange, and proposes an approach to identify the effects of important factors on the dynamic characteristics of the market prices by constructing sparse models. The concept of the sparse modeling is an attractive approach to identify the effects of variables mechanically based on the coefficients in constructed models. Additionally, the proposed approach selects variables of particular importance based on statistical testing using the coefficients. The proposed scheme was applied to a real-world price dataset and discussed in the context of representation errors and interpretability. The results show that the proposed scheme is an effective approach for representing the impact of variables on electricity market price.