An Ontology-Based Information Extraction System for Organic Farming

Introduction

Socio-economic developments have increased rapidly in recent times due to the increase in the adoption of Information and Communication Technology (ICT) by governments, profit and non-profit institutions. This has led to positive changes in livelihoods and growth of information and knowledge societies. As modern ICT and related technologies converge, the ability to produce, access, adapt and apply information has increased tremendously in the newly emerging information and knowledge societies (Okello et al., 2013). The need for access to accurate, relevant and timely information for agricultural development cannot be overemphasized (Mojaki and Keregero, 2019). Farmers in many developed countries have adopted precision agriculture to improve agricultural productivity (Patel and Seyyed, 2014) through the extensive usage of ICT in weather forecasting, farming equipment, irrigation, soil analysis, etc. (FAO, 2017). However, not much concern has been given to acquiring, storing and making organic farming-based information available in a concise manner to the current and intending software developer who may wish to develop applications for farmers.

Organic farming is a modern farming approach that encourages the use of biological practices and management methods which improves soil quality through increased soil organic matter, cycle nutrients, water and carbon absorption/storage. It optimizes the productivity and fitness of different communities within the agro-ecosystem, including plants, people, livestock and soil organisms (OMAFRA, 2019). Francis (2015) defines organic farming as “the production of the crop, animal, and other products without the use of synthetic chemical fertilizers and pesticides, transgenic species, or antibiotics and growth-enhancing steroids, or other chemicals”. Organic agriculture aims to create an environmental and economic viable agricultural system that is integrated and humane with optimum dependence on renewable resources mostly from the farm (Deshmukh and Babar, 2013). Practicing organic farming helps in ensuring sustainability in agricultural practices, thus, easy access to relevant and concise information will offer tremendous benefits to farmers in optimizing their productivity and ensuring sustainability for the future generation.

Although access to existing information sources have been simplified lately due to growth in digital library and increased access to the Internet, the task of finding, extracting and aggregating relevant information from this pool of data is becoming increasingly difficult (Jiang et al., 2020; Ojokoh and Ayokunle, 2012). Information Extraction (IE) is a challenging task due to the complexity and ambiguity of natural language. For example, the same fact could be expressed in several forms using different sentences across multiple documents, or even knowledge repositories (Piskorski and Yangarber, 2013; Gutierrez et al., 2016). Such data are transmitted in various forms as structured, unstructured, free-text documents, multi-media files, images, videos, etc., which makes it hard to search or analyze (Piskorski and Yangarber, 2013; Yang and Zhu, 2011). Hence, there is a growing need for an efficient and effective technique for retrieving structured information from free-text data in order to discover valuable and relevant knowledge. Information Extraction (IE) is defined as “identifying a predefined set of concepts in a specific domain, ignoring other irrelevant information, where a domain consists of a corpus of texts having a specified information need. It involves deriving structured factual information from the unstructured text” Chang et al. (2006). It helps users to query the required information from a large pool of useful data. The use of ontology for formal and explicit specification domain concepts has been helpful in IE, making Ontology-Based Information Extraction (OBIE) a clear sub-discipline of knowledge extraction (Wimalasuriya and Dou, 2010). Information retrieval does not require a specific algorithm or a method; it is based on ontology, which is used to support and guide algorithms for efficient and relevant IE (Alejandria & Vinluan, 2017).