Extraction of Plant Identification Keys Using Approximate String Matching for Species Properties Classification

摘要

Most biologists keep data in separate databases. These databases are not necessary well-structured. Plant identification keys are among such data. They are data-rich description containing plant identification terminologies and maybe used to identify various plant species. The way the data is kept often requires the species identification to be done using rules that are applied sequentially. Done manually, this is very time consuming. Information extraction (IE) is a process of selecting information such as names, terms, or phrases, from a natural language text documents. This information is then structured into a specified template for retrieval. This method is applied to plant identification keys kept by the biologists. Before the keys are extracted from the description, they have to go through a number of processes. In this paper, we illustrate the pre-processing and processing methods with an example from a database, with emphasis on the approximate string matching algorithm to extract the most relevant keys from the description.