Predictive data mining is an important research direction in big data engineering. It has extensive demand in the fields of electric power, transportation, mining, agriculture, meteorology, etc. For example, the prediction of power load and system vulnerability in the power system, the prediction of traffic passenger flow in the transportation system, the prediction of mine pressure, groundwater and harmful gas emission in the mining system, the prediction of water and soil meteorological relations and diseases and pests in the agricultural system, and the prediction of temperature, humidity, air pressure, rain and snow in the meteorological system. Early data prediction and analysis mostly used linear prediction systems, using a variety of acquisition conditions and a variety of pre-processing methods of data (linear equation superposition, linear space adjustment). After obtaining the linear law of data, the forward prediction of the data law function was realized.

However, there are two systematic problems in linear programming. One is that the sensitivity of data forward pushing amount is closely related to the total data amount. Generally, the data forward pushing amount cannot exceed 10% of the total data amount, otherwise it will lead to the rapid decline of prediction sensitivity. The other is that the linear prediction algorithm cannot reflect the periodic law of data and the above-mentioned power, transportation, mining, and agriculture. The data in meteorology and other fields are highly cyclical, and this systematic problem seriously affects the prediction sensitivity at the data inflection point. The subject introduces nonlinear functions into the estimation of data prediction curve, especially trigonometric function and Fourier analysis. The wavelet analysis algorithm is used to reduce the data noise. The difference between the original data and the data after wavelet analysis is calculated to extract the noise data, the Fourier analysis is used to extract the characteristic matrix, and the trigonometric function regression algorithm is used to obtain the final prediction curve. The algorithm has realized laboratory simulation in many fields, and the subject will strengthen its field practical research.

This Special Issue welcomes original research and review articles focused on nonlinear prediction systems for data from a complex system.

Potential topics include but are not limited to the following:

• Nonlinear data prediction of power system vulnerability
• Nonlinear data prediction of power system load dispatching
• Nonlinear data prediction of common equipment faults in power system
• Nonlinear data prediction of passenger flow of Urban Rail Transit
• Urban public transport transfer volume driven scheduling plan selection strategy nonlinear data prediction
• Nonlinear data prediction of periodic ground pressure and rockburst disaster in deep mines
• Nonlinear data prediction of water inflow and harmful gas emission in deep mines
• Nonlinear data prediction of agricultural disease and pest risk
• Nonlinear data prediction of agricultural output and transaction price of supply and demand
• Nonlinear data prediction of extreme weather and related natural disasters
• Long period data nonlinear data prediction of sunshine and temperature