A Hybrid Approach for Predicting Dissolved Oxygen in Aquaculture to Enhance Water Quality

  • Unique Paper ID: 167865
  • Volume: 11
  • Issue: 4
  • PageNo: 510-516
  • Abstract:
  • This study proposes a hybrid model, combining Light Gradient Boosting Machine (LightGBM) and Bidirectional Simple Recurrent Unit (BiSRU), to predict dissolved oxygen (DO) levels in aquaculture environments accurately and efficiently. Initially, linear interpolation and smoothing techniques are employed to identify significant parameters, followed by LightGBM algorithm's utilization to determine the relevance of dissolved oxygen and predict its levels in intensive aquaculture settings. Furthermore, an attention mechanism is implemented to assign varying weights to BiSRU's hidden states, enhancing its predictive capabilities. The model demonstrates remarkable performance, accurately anticipating DO fluctuations over a 10-day period in just 122 seconds, with an impressive accuracy rate of 96.28%. This approach addresses the challenges faced by traditional methods in predicting DO levels due to the nonlinear, dynamic, and complex nature of aquatic environments. The significance of maintaining water quality in aquaculture for optimal productivity underscores the importance of accurate DO prediction. Overall, the proposed hybrid model offers a promising solution for enhancing DO prediction accuracy and speed, contributing to effective disease prevention and economic sustainability in aquaculture operations.

Cite This Article

  • ISSN: 2349-6002
  • Volume: 11
  • Issue: 4
  • PageNo: 510-516

A Hybrid Approach for Predicting Dissolved Oxygen in Aquaculture to Enhance Water Quality

Related Articles