Machine Learning Model to Predict Potential Fishing Zone

Mudliar, Swaroop Laxmi; Shashank, Sai ; Chandak, M


A key challenge today in aquatic environment conservation is the  accurate tracking of the spatial distribution of various human impacts on activities like fishing. In the present paper an approach to identify the potential fishing zones in deep sea waters is developed using Autoregressive Integrated Moving Average(ARIMA) and Random Forest model. A large data set containing Indian fishing vessel track from 2017-2019 was taken as database.  In the present paper an approach a methodology was developed to detect and map fishing activities. Validation of the model was done against expert label datasets which showed detection accuracy of 98 percent. Our study represents the first comprehensive approach to detect and Identify Potential fishing zones with the help of two important water quality indicators viz Dissolved Oxygen and Salinity.


Water quality indicators; Satellite-based Automatic Information Systems (S-AIS); Datamining; Potential fishing zone (PFZ)

Full Text: PDF (downloaded 1423 times)


  • There are currently no refbacks.
This abstract viewed 1646 times