CocoPest Watch System
A GIS-based pest infestation monitoring and forecasting system using standard ARIMA(1,1,1) time-series analysis to predict coconut pest outbreaks in Aurora, Zamboanga del Sur.
What is the ARIMA Forecasting Model?
ARIMA (Auto-Regressive Integrated Moving Average) is a statistical model designed for time-series forecasting. It analyzes past observations to identify momentum, stationarity, and error patterns.
Auto-Regressive (AR - p=1)
Uses past differenced momentum (φ₁=0.65) to project the next period change rate.
Integrated (I - d=1)
Applies 1st-order differencing (ΔY_t = Y_t - Y_t-1) to make data stationary.
Moving Average (MA - q)
Corrects for prior step prediction errors (θ₁=0.30) to prevent trend divergence.
Actual Dataset Sample Calculation (Barangay Acad)
System Features
Interactive map visualization showing infestation levels across barangays with color-coded severity indicators. Click any barangay to see detailed infestation data.
Uses the standard ARIMA(1,1,1) time-series model to predict pest infestation trends based on differenced historical observations and 95% confidence intervals.
Comprehensive statistical evaluation including 95% confidence prediction envelopes (lower and upper bounds) to account for environmental uncertainty.
Full support for Microsoft Excel (.xlsx) file format and instant PDF Model Specification export for official reporting and offline documentation.
Full CRUD (Create, Read, Update, Delete) operations for all data entities - Barangays, Insect Pests, and Infestation Reports through an intuitive admin interface.
Dashboard updates in real-time as new infestation data is recorded. Monitor current infestation status and historical trends at a glance.
How the System Works
Field reporters submit infestation data including barangay location, pest type, number of infested trees, and date of report.
First-order differencing (ΔY_t = Y_t - Y_t-1) is applied to eliminate baseline drift and establish stationary baseline change rates.
The standard ARIMA(1,1,1) model applies AR(1) momentum (φ₁=0.65) and MA(1) error corrections (θ₁=0.30) to calculate multi-step predictions.
Results are projected onto interactive GIS maps and charts, with one-click export to publication-ready PDF documentation.