Predicting the oriental fruit fly Bactrocera dorsalis (Diptera: Tephritidae) trap catch using artificial neural networks: a case study

Abstract

The oriental fruit fly Bactrocera dorsalis (Hendel) is a very serious pest of fruit trees, causing enormous economic losses globally. The present study examines the capability of an artificial neural network (ANN) with a Quasi-Newton (QN) algorithm to predict a fruit fly trap catch and compare the results with those of a traditional regression model. MATLAB 7.0 was used to develop ANN programming and the fortnightly measurement of 14 input variables (abiotic along with biotic variables) provided the database for analysing the ANN model. An input model using a total of 14 identified input nodes with a selected QN-ANN structure (14-25-20-1) gave an optimum result. In general, the present study showed that an ANN could be used to estimate fruit fly trap catch with enhanced accuracy (R 2 ¼ 0.92; root mean square error (RMSE) ¼ 23.75; Nash -Sutcliffe efficiencies ¼ 0.99) over traditional regression models (R 2 ¼ 0.76; RMSE ¼ 30.28; Nash -Sutcliffe efficiencies ¼ 0.76). This finding helps the region-specific fruit fly monitoring and management programmes that lack long-term historic data.

Predicting Rice Pest Population Occurrence with Satellite-Derived Crop Phenology, Ground Meteorological Observation, and Machine Learning: A Case Study for the Central Plain of Thailand

“…As has been observed in previous studies [45], our findings indicated that the ANN model yielded more accurate pest prediction than the MLR model. This may be because the weights assigned to each neural network connection allow a more precise pattern in the input data to be identified [44]. The RF model also performed better than the MLR model, but not better than the ANN model for our dataset.…”

“…ANN has been successfully applied in various disciplines, such as agriculture, medical science, education, finance, management, security, engineering, trading commodity, and art [33][34][35][36][37][38][39][40][41][42]. For pest forecasting, ANN has been applied to model the pest population dynamics for the pod borer Helicoverpa armigera (Lepidoptera: Noctuidae) [43], the fruit fly Bactrocera dorsalis (Diptera: Tephritidae) [44], rice pests [45] and the paddy stem borer (Scirpophaga incertulas) [46]. In most of these studies, the authors referred to the limitation of linear multiple regression (MLR) and concluded that the ANN technique is more precise than MLR and is a good choice for predicting pest occurrence.…”

The Fallacy of Year-Round Breeding in Polyphagous Tropical Fruit Flies (Diptera: Tephritidae): Evidence for a Seasonal Reproductive Arrestment in Bactrocera Species

“…Over the last decade numerous researchers have applied increasingly sophisticated multivariate, neural network and fuzzy-logic analyses to seek correlations between B. dorsalis population numbers and weather and/or crop variables [ 25 , 26 , 27 , 120 , 121 , 122 , 123 , 124 , 125 ]. In some cases correlations have been very high and it could be argued that such analyses are evidence against B. dorsalis having anything more than a ‘simple’ weather/host-availability driven phenological cycle.…”

“…Thus these studies show that fly populations are correlated with the crop still to come, rather than being directly correlated with breeding within that crop, a subtle but important difference that reinforces the early work of Newell and Haramoto [ 116 ] in understanding the difference between adult abundance and adult reproduction. Finally, as some of these models use ‘hidden layer’ components and are only tested against internal data [ 26 , 27 , 121 ], evaluating their predictive capacity against independent data sets is not possible.…”