Elevating Forestry Prediction: A Study on Machine Learning Model for Plantations Survival Rate Analysis

Abstract

This paper details the development and preliminary findings of a machine learning model designed to predict the survival rate of plantations. Drawing data from official sources, various vegetation indices were used as features for the predictive model. Initial results show potential, despite certain limitations, suggesting avenues for further enhancement and application.

Introduction

I. INTRODUCTION

Plantations play a significant role in environmental conservation and economic sustenance. Predicting their survival rates becomes essential for sustainable development and forest management. With advancements in remote sensing and machine learning, this research aims to develop a predictive model using satellite imagery indices and the Gradient Boosted Trees algorithm to determine the survival rate of plantations.

II. STUDY AREA

The study area includes plantations done by MP Forest Department in East Chhindwara Division from 2015 to 2018.

III. METHODOLOGY

A Data Collection

The primary source for the research data was the Madhya Pradesh Forest Department's official portal www.mpforest.gov.in, from which the Plantation Survival Report and KMLs of plantations were extracted.

IV. RESULTS AND INTERPRETATION

Of the 112 plantations analysed, the model accurately predicted the survival of 91. The overall accuracy stood at 81.25%. However, the research encountered a limitation in the form of the database's restricted scope, sourced from just one forest division spanning four years only2015 to2018. Consequently, predictions for plantations with reduced survival rates showed significant errors.

V. DISCUSSION

The model, in its present iteration, holds potential, even if the accuracy isn't at an optimal level. Its primary value lies in assisting field officers in identifying plantations at risk. By flagging potential failures, proactive measures can be initiated to mitigate issues. For future iterations, it's imperative to diversify and expand the dataset. Incorporating additional indices and geometric features could further enhance the model's predictive capabilities.

VI. FUTURE WORK

1. Augmenting the dataset is a priority, ensuring diverse data sources to refine the model further.
2. New indices and geometric parameters, especially features like land surface temperature, will be considered in the updated model.
3. Plans to automate the entire model are underway using platforms such as Google Colab, making the process more user-friendly.

References

[1] Liu, Pei. Machine-Learning-Based-Survival-Analysis. github.com/liupei101 [2] Elith, J., & Leathwick, J.R. (2008). A working guide to boosted regression trees. besjournals.onlinelibrary.wiley.com [3] Arain, A. (Year not specified). Machine Learning Approach to Quantify Leaf Depletion. github.com/Arain23

Copyright

Copyright © 2024 Akshaya Rathore. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.