Dynamic Evolution Mechanism of Digital Entrepreneurship Ecosystem Based on Text Sentiment Computing Analysis

Area of Science:

• Digital Entrepreneurship
• Computational Social Science
• Ecosystem Dynamics

Background:

• Current entrepreneurial ecosystems face limitations.
• Research into digital entrepreneurial ecosystems is crucial for addressing these gaps.
• Understanding the dynamic evolution mechanisms is key to fostering innovation.

Purpose of the Study:

• To investigate the dynamic evolution mechanism of the digital entrepreneurship ecosystem.
• To propose and evaluate an improved text sentiment computing analysis model for this purpose.
• To predict and analyze the future development of digital entrepreneurial ecosystems.

Main Methods:

• Development of an improved Bi-directional long short-term memory (Bi-LSTM) model incorporating a multilayer neural network.
• Utilizing an optimized Naive Bayes algorithm with Euclidean distance weighting for enhanced classification.
• Sentiment analysis of text data to capture ecosystem dynamics.

Main Results:

• The improved Bi-LSTM model demonstrated higher accuracy (1.1% increase) and F1 value (0.6% increase) compared to traditional Bi-LSTM.
• The model showed significant improvements over traditional Cellular Neural Network (CNN) algorithms, with accuracy increasing by 4%, recall by 14%, and F1 value by 9%.
• Running on the Spark platform improved runtime by 320% with a minor accuracy trade-off.

Conclusions:

• The proposed improved Bi-LSTM model effectively analyzes and predicts the dynamic evolution of digital entrepreneurship ecosystems.
• The model's strong non-linear fitting ability is validated by performance improvements on large datasets.
• The research confirms the feasibility and potential of digital entrepreneurship ecosystems for future development.