Generating fractal functions associated with Suzuki iterated function systems

M. V. Patel, Geetika Verma, Andrew Eberhard, Archana RaoarXiv 2025 · DOI · arXiv

Abstract: This article constructs a fractal interpolation function, also referred to as $\alpha$-fractal function, using Suzuki-type generalized $\varphi$-contraction mappings (STGPC). The STGPC is a generalization of $\varphi$-contraction mappings. The process of constructing $\alpha$-fractal functions using the STGPC is detailed, and examples of STGPC are given. The FIF has broad applications in data analysis, finance and price prediction. We have included a case study analyzing the price volatility of spinach in the Azadpur vegetable market in New Delhi. The fractal analysis gives a unique perspective on understanding price fluctuations over a period. Finally, the box-dimensional analysis is presented to comprehend the complexity of price fluctuations.

Introduction

Traditional interpolation methods, such as linear or polynomial curves, are effective for well-behaved, smooth data. However, they often fail to capture the inherent irregularity and "roughness" of real-world phenomena like stock prices, weather patterns, or physiological signals. Fractal Interpolation Functions (FIFs) address this by introducing self-similarity and intentional complexity between data points.

Historically, the Iterated Function Systems (IFS) used to generate these fractals have relied on Banach contraction mappings, which impose strict mathematical constraints. This research introduces a novel method for constructing $\alpha$-fractal interpolation functions by utilizing Suzuki-type generalized $\varphi$-contraction mappings (STGPC). By relaxing these traditional requirements, the proposed approach offers greater flexibility in modeling highly volatile and irregular datasets.

Methodology

To build robust fractal models, the authors move beyond standard contraction mappings. While a standard contraction mapping strictly requires the distance between any two points to shrink after the mapping is applied, the Suzuki condition is less restrictive. It only requires the shrinking condition to hold if a specific preliminary distance condition is met.

The authors mathematically prove that an IFS built with STGPC still guarantees a unique "attractor," which forms the resulting fractal curve. They construct an $\alpha$-fractal function where $\alpha$ acts as a vertical scaling factor—essentially a "volatility dial." Setting $\alpha = 0$ yields a classical smooth interpolation, while increasing $\alpha$ injects varying degrees of chaos into the model.

Table of cases defining the mapping transformations used in the fractal construction.

As shown in the mapping examples above, the framework allows for complex, piecewise definitions that can be tuned to specific data characteristics. To quantify the complexity of the resulting curves, the authors employ Box-Counting Dimension ($D$). This metric provides a numerical value for the "roughness" of the data, measuring how detail changes as the scale of observation shifts.

Case Study: Spinach Price Volatility

The theoretical framework was applied to a real-world economic dataset: average monthly price data for spinach (in ₹/kg) from the Azadpur vegetable market in New Delhi, India, spanning September 2023 to July 2024. Agricultural markets are notoriously volatile due to seasonal shifts and supply chain issues, making them ideal for fractal modeling.

By treating historical price data as a fractal rather than a standard time-series curve, the researchers captured intricate, micro-level details of price changes. The study demonstrated that the mathematical results align with visual intuition: a higher scaling factor ($\alpha$) results in a higher box dimension ($D$), directly quantifying a higher degree of complexity in the modeled price trend.

Key Results

  • Theoretical Foundation: The authors successfully established the Read-Bajaraktarevic (RB) operator using STGPC on a complete metric space, proving the existence and uniqueness of the resulting FIF.
  • Quantitative Complexity: The box dimension ($D$) of the spinach price data was calculated under different scaling factors:
  • For $\alpha = 0.4$, $D \approx 1.60$.
  • For $\alpha = 0.6$, $D \approx 1.77$.
  • Using a variable vector of $\alpha$ values across different sub-intervals yielded $D \approx 1.41$.

Conclusion and Future Work

This work proves that strict continuity or standard $\varphi$-contractions are not strictly necessary to find a fixed point in fractal geometry. The STGPC approach provides a robust alternative for modeling volatility.

While this study utilized a relatively small dataset of 11 monthly averages, future research could apply this method to larger financial contexts, such as stock market indices or broader economic indicators. Additionally, there is significant potential in using STGPC-based fractal methods not just for analyzing historical volatility, but for actively forecasting future market trends.

Summary updated

Abstract

This article constructs a fractal interpolation function, also referred to as $α$-fractal function, using Suzuki-type generalized $φ$-contraction mappings (STGPC). The STGPC is a generalization of $φ$-contraction mappings. The process of constructing $α$-fractal functions using the STGPC is detailed, and examples of STGPC are given. The FIF has broad applications in data analysis, finance and price prediction. We have included a case study analyzing the price volatility of spinach in the Azadpur vegetable market in New Delhi. The fractal analysis gives a unique perspective on understanding price fluctuations over a period. Finally, the box-dimensional analysis is presented to comprehend the complexity of price fluctuations.
BibTeX
@misc{patel2025generating,
  author = {M. V. Patel and Geetika Verma and Andrew Eberhard and Archana Rao},
  title = {{Generating fractal functions associated with Suzuki iterated function systems}},
  year = {2025},
  eprint = {2501.16851},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2501.16851},
  howpublished = {arXiv},
}
Figure