Interpolation And Forecasting Of Data In The Big Data Era

Reliable data have always played a vital role in time series analysis and research. Nevertheless, missing data, which can bias the original properties of the time series if the pattern of missed data is systematic, are also a very common phenomenon in observed processes. Thus, how to address missing values is a very important challenge, especially in the upcoming big data era. This paper proposes the Cuckoo Search-Transforming Fractal Iteration Functions (CS-TFIFs) method and the CS-TFIFs-Winners Combining (CS-TFIFs-WC) method for interpolating missing values. The former method skilfully transforms classical Fractal Iteration Functions to make it possible to calculate a specified point’s missing value, which is difficult to obtain with the traditional approach. Then, to optimise the parameter transforming process that is generated, the cuckoo search algorithm (CS) and TFIFs are synthesised into a novel model, CS-TFIFs. Considering classical interpolation methods, such as Linear, Cubic Spline and Piecewise Cubic Hermite Interpolation Polynomial (PCHIP), having some born advantages, the inspiration of winners combining is sparked. Here, the winners are the top 3 models in terms of the interpolation accuracy, which is measured by the mean absolute percentage error (MAPE). Additionally, the CS algorithm is used to obtain the best weights of these 3 models in this combined model, CS-TFIFs-WC. In electricity markets, forecasting systems for electricity prices have been always played an important role; however, accurate price forecasting, which is critical for both government regulatory agencies and public power companies to make decisions, remains a challenging problem for this type of system. Accurate price forecasting has also become more important in recent years due to our society becoming heavily reliant on electricity. Based on the Index of Bad Samples Matrix (IBSM), a novel method including unsupervised learning, the Optimization Algorithm (OA), DCANN and Updated DCANN, which is a hybrid system of supervised and unsupervised learning, are proposed in this paper for forecasting day ahead electricity price. This model creatively applies the idea of deleting bad samples and searching quality inputs to develop and learn, thus creating a dynamic structure.