A Bacterial Foraging Optimization Algorithm Based on Normal Distribution for Crowdfunding Project Outcome Prediction
| Autor: | Shuang Geng, Yingsi Tan, Shilian Chen |
|---|---|
| Rok vydání: | 2021 |
| Předmět: |
education.field_of_study
Computer science business.industry Population Evolutionary algorithm Initialization Swarm behaviour Particle swarm optimization Machine learning computer.software_genre Normal distribution Artificial intelligence Gradient boosting Evolution strategy business education computer |
| Zdroj: | Lecture Notes in Computer Science ISBN: 9783030787424 ICSI (1) |
| DOI: | 10.1007/978-3-030-78743-1_46 |
| Popis: | Crowdfunding is a concept that raising fund for different individual or organization to conduct creative projects and it has gained more and more popularity during these years. Fund used for projects can reach to billions of dollars, so it’s very significant to perfectly predict multiple crowdfunding ads. To improve the accuracy of crowdfunding project outcome prediction, a modified Bacterial Foraging Optimization Algorithm (NBFO) through population initialization, reproduction and elimination-dispersion was proposed to cooperate with Light Gradient Boosting Machine (LightGBM). This paper used normal distribution through the period of population initialization and elimination-dispersion. Moreover, during reproduction, selective probability was introduced to enhance the performance of bacteria. Experiments used 5561 valid data collected from Kickstarter from June 2017 to February 2018, and compared the predictive power of LightGBM incorporated with Particle Swarm Optimization (PSO), Bee Colony Optimization (BCO) and Evolutionary Strategy (ES). Results showed that the performance of NBFO surpasses all comparative algorithms. The performance of LightGBM incorporated with other swarm intelligent algorithms and evolutionary algorithm are discussed. Findings in this study contribute to the study of crowdfunding, Light Gradient Boosting Machine, swarm intelligent algorithm and evolutionary algorithm. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|