Document worth reading: “A Comparative Survey of Recent Natural Language Interfaces for Databases”

Over the last few years pure language interfaces (NLI) for databases have gained necessary traction every in academia and commerce. These applications use very completely totally different approaches as described in newest survey papers. However, these applications have not been systematically in distinction in opposition to a set of benchmark questions in an effort to scrupulously think about their functionalities and expressive vitality. In this paper, we give an abstract over 24 not too way back developed NLIs for databases. Each of the applications is evaluated using a curated guidelines of ten sample questions to point their strengths and weaknesses. We categorize the NLIs into 4 groups primarily based totally on the methodology they’re using: keyword-, pattern-, parsing-, and grammar-based NLI. Overall, we realized that keyword-based applications are adequate to answer simple questions. To clear up further superior questions involving subqueries, the system desires to make use of some sort of parsing to find out structural dependencies. Grammar-based applications are common basically essentially the most extremely efficient ones, nonetheless are extraordinarily relying on their manually designed tips. In addition to providing a scientific analysis of a very powerful applications, we derive lessons realized which could be necessary for designing NLIs that will reply a range of particular person questions. A Comparative Survey of Recent Natural Language Interfaces for Databases