Paper: arXiv 2401.05441

Authors: Ali Mehrban, Pegah Ahadian

Abstract

This paper describes an architecture for predicting the price of cryptocurrencies for the next seven days using the Adaptive Network Based Fuzzy Inference System (ANFIS). Historical data of cryptocurrencies and indexes that are considered are Bitcoin (BTC), Ethereum (ETH), Bitcoin Dominance (BTC.D), and Ethereum Dominance (ETH.D) in a daily timeframe. The methods used to teach the data are hybrid and backpropagation algorithms, as well as grid partition, subtractive clustering, and Fuzzy C-means clustering (FCM) algorithms, which are used in data clustering. The architectural performance designed in this paper has been compared with different inputs and neural network models in terms of statistical evaluation criteria. Finally, the proposed method can predict the price of digital currencies in a short time.

Complexity vs Empirical Score

  • Math Complexity: 7.0/10
  • Empirical Rigor: 5.0/10
  • Quadrant: Holy Grail — high math complexity, high empirical rigor

Why this score: The paper employs advanced adaptive network-based fuzzy inference systems (ANFIS) with hybrid learning and clustering algorithms, indicating high mathematical density, while the empirical rigor is moderate due to the use of statistical evaluation criteria (RMSE, MAE) on historical cryptocurrency data, though it lacks explicit backtesting or live trading implementation details.

Research Flowchart

  flowchart TD
  A["Research Goal:<br>Predict Cryptocurrency Prices<br>for Next 7 Days"] --> B["Input Data:<br>BTC, ETH, BTC.D, ETH.D<br>(Daily Timeframe)"]
  B --> C["Data Clustering:<br>Grid Partition /<br>Subtractive / FCM"]
  C --> D["Model Architecture:<br>ANFIS<br>Adaptive Neuro-Fuzzy<br>Inference System"]
  D --> E["Training Algorithms:<br>Hybrid &<br>Backpropagation"]
  E --> F["Evaluation:<br>Comparison with<br>Neural Networks & Inputs"]
  F --> G["Outcome:<br>Accurate Short-Term<br>Price Forecasting"]

Evidence Card

What the authors claim, on what data, and who has checked it. Read from the paper's own text by a language model: every quote is verbatim, nothing was re-run. How to read this card

Proposes

The paper proposes an ANFIS-based architecture to forecast cryptocurrency prices for the next seven days.

Source sentence
This paper describes an architecture for predicting the price of cryptocurrencies for the next seven days using the Adaptive Network Based Fuzzy Inference System (ANFIS).

Assumes 5

  • Historical price data is treated as a time‑series suitable for forecasting. data
    Source sentence
    This study shows that historical price data is a kind of time series data, so we can use time series methods to forecast prices.
  • 90% of the data is used for training and 10% for testing. data
    Source sentence
    In this study, we consider 90% of the data for training and 10% for testing.
  • Data are collected on a daily candle (daily timeframe). data
    Source sentence
    This paper analyzes the historical data of BTC and ETH and BTC.D and ETH.D at the end of a daily candle.
  • The BTC.D dominance index is relevant for predicting cryptocurrency prices. model
    Source sentence
    Therefore, using this index to predict the price of cryptocurrencies can be appropriate.
Show 1 more
  • Tuning ANFIS parameters (e.g., clustering, training algorithms) improves forecasting evaluation criteria. model
    Source sentence
    We first need to tune and assess the parameters of the ANFIS system to see which ones enhance the evaluation criteria for price forecasting.

Data

  • BTC, ETH, BTC.D, ETH.D daily price data · Tradingview market-prices public
    August 17, 2017 to June 3, 2022 · Bitcoin, Ethereum and their dominance indices
    Source sentence
    This paper analyzes the historical data of BTC and ETH and BTC.D and ETH.D at the end of a daily candle. This historical data has been collected from August 17, 2017, to June 3, 2022, from Tradingview.

Recognised sources: Kraken

Costs

Not applicable The paper reports no trading or execution result.

Finds 5

  • ANFIS improves RMSE and RMSRE when using FCM clustering and backpropagation training. predictive-accuracy
    Source sentence
    Based on Table 2, we see that the ANFIS system improves the evaluation criteria such as RMSE and RMSRE by using the FCM algorithm for data clustering and the backpropagation algorithm for data training.
  • Including BTC.D as an input improves evaluation criteria compared to using only BTC price. predictive-accuracy
    Source sentence
    We test the system with two inputs, the price of BTC and the value of BTC.D, and based on Table 2, the evaluation criteria improve compared to using only our BTC historical data for forecasting.
  • Including ETH.D as an input improves evaluation criteria compared to using only ETH price. predictive-accuracy
    Source sentence
    We do the same for ETH, and based on Table 3, the evaluation criteria improve when we use ETH.D along with the price of ETH.
  • ANFIS outperforms ANN in RMSE for BTC prediction (both training and test), while ANN performs better for ETH test data. performance
    Source sentence
    In Table 4, we contrast this method with the ANN method and see that the ANFIS method performs better in terms of RMSE for BTC prediction in both training and test data, but the ANN method does better for ETH prediction in test data.
  • The forecasted price over seven cycles differs only slightly from the actual price in each cycle. predictive-accuracy
    Source sentence
    Finally, based on Fig. 4, the forecasted price differs slightly from the actual price in each cycle.

Checked by others

No other paper in the archive has been found to support or contradict these findings (compared with 17 related papers). Coverage is partial, so this is neither replication nor refutation.

Code

No repository by the authors found in the paper.

Extracted 2026-11-14 · same card as JSON · A language model chose these items: check the source sentence before relying on one.