Machine learning model(s) for pest management strategies

People

External Member

Khandaker Asif Ahmed, CSIRO (https://bit.ly/2KXPWZU)

Description

Proper identification and classification of insect pest is a crucial component for integrated pest management strategies, currently getting popularized worldwide as pesticide-free insect control strategy. Classification taxonomic identification is a cumbersome process, which need insect experts with vast knowledge of insect morphology but also human-prone due to species complexity. Currently, genomic features (publicly available on NCBI database in limited amount) based approach is an alternative way to solve insect identification and taxonomy problems. 

Goals

The project aims to solve the paradigm of insect identification and taxonomy by machine learning approach to-

  1. Extract morphological and morpho-metric data from publicly available datasets (e.g. -ANIC, OZCAM, GBIF, livingAustralia)
  2. Develop algorithm to place closely related insects in common clusters and validation using DNA-sequence data.
  3. Analyze ancestry among different species.

The data, trained from current project, will be utilized to identify insect species from museum or life samples and monitor prevalence of insect pests in agricultural fields.

 

Updated:  1 June 2019/Responsible Officer:  Dean, CECS/Page Contact:  CECS Marketing