Scholarlink

<p>The authors present two distinct approaches for forecasting time series of complex systems characterized by non-stationarity and uncertainty. The first approach is a fuzzy logic-based method, which incorporates expert opinion to define the system using membership functions and fuzzy decision rules. This approach facilitates the development of knowledge applications capable of processing ambiguity and inaccuracy. The second approach is a statistical method that captures the interdependencies among multiple time series. The objective is to introduce valid procedures suitable for predicting economic time series for management purposes. Prediction models are developed and validated through a case study that forecasts Tesla electric vehicle sales for the four quarters of 2024 using historical data on variables such as price, ownership costs, range, and sales. By comparing the results with expert estimates and contemporary knowledge, the authors explore the added value of expert judgment in a predictive algorithm based on fuzzy logic compared to the statistical VAR method.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Predictive Power of Fuzzy Model vs. Statistical Model: Prediction of Tesla Car Sales

  • 期刊论文

  • 发表日期：

    2025年10月22日

查看全文
  • Simona Hašková,
  • Petr Šuleř,
  • Martin Smrt

摘要

The authors present two distinct approaches for forecasting time series of complex systems characterized by non-stationarity and uncertainty. The first approach is a fuzzy logic-based method, which incorporates expert opinion to define the system using membership functions and fuzzy decision rules. This approach facilitates the development of knowledge applications capable of processing ambiguity and inaccuracy. The second approach is a statistical method that captures the interdependencies among multiple time series. The objective is to introduce valid procedures suitable for predicting economic time series for management purposes. Prediction models are developed and validated through a case study that forecasts Tesla electric vehicle sales for the four quarters of 2024 using historical data on variables such as price, ownership costs, range, and sales. By comparing the results with expert estimates and contemporary knowledge, the authors explore the added value of expert judgment in a predictive algorithm based on fuzzy logic compared to the statistical VAR method.