this is a very basic question.
For my master thesis I need to estimate the power consumption for the current month. I have a lot of historical data of the power consumption. I have data for every 15min of the last 3 years. What I want to show is for example the power consumption of the last 3 months and next to it the estimated consumption for the current month. I could think of a lot of ways to calculate this, but I thought there must be some basics and literature I could use to find a good solution.
Any recommendations on this topic? I mustn't be too complicated thought =)
It seems you want to estimate current month consumption based solely on consumption in previous 3 months. Other information such as time of year, weather etc. may improve estimate. But if you use only past consumption, then you could run a regression with dependent variable as a month's consumption and the independent variable as the consumption in the previous 3 months. However, you will get only 9 observations with nonoverlapping 4-month windows. Time series models will be more powerful and will allow you to use more observations such as 36 monthly consumption values. ARMA models may work well. A classic reference is Time Series Analysis by Hamilton. See R command documentation.