Multifractal time series analysis using the improved 0–1 test model

• We develop a new method of MSFA to analyze multifractal properties of time series.
• We investigate multifractality of the different signals, using the MSFA and MF-DFA.
• MSFA shows diverse sensitivity for the parameters than MF-DFA to multifractality.

In this paper, we propose a new method of multifractal signal fluctuation analysis based on the 0–1 test that has previously been used for determining whether a time series is deterministically chaotic or not. It is shown that the method works well both on artificial and actual data, and tests suggest several advantages over only the 0–1 test. Moreover, analysis based on descriptive statistics is provided to obtain the inherited multifractality of nonstationary time series. Especially, the multifractal signal fluctuation analysis technique is validated to be appropriate for traffic flow data.