A state-of-the-art survey on recommendation system and prospective extensions

• Published in: Computers and electronics in agriculture Vol. 178; p. 105779
• Main Authors: Patel, Krupa, Patel, Hiren B.
• Format: Journal Article
• Language: English
• Published: Amsterdam Elsevier B.V 01.11.2020; Elsevier BV
• Subjects: Agriculture; Algorithms; Collaborative filtering; Content-base; Crop recommendation system; Crops; Domains; Parameters; Pesticides; Recommendation algorithms; Recommender systems; Statistical analysis; Systems analysis; Content-base; Agriculture; Collaborative filtering; Crop recommendation system; Recommendation algorithms
• ISSN: 0168-1699; 1872-7107
• DOI: 10.1016/j.compag.2020.105779

•Recommendation system is new era of research to predict things to end user.•Content-based and collaborative filtering are two main building block of recommendation system.•Recommendation system suffers from cold-start, sparsity and shilling attack problems.•Different online and offline dimensions are available to evaluate recommendation system.•Recommending crops, pesticides, fertilizer in agriculture based on soil properties can help farmers to gain profit. With the new era of the Internet, we have a large amount of data available in the form of ratings, reviews, graphs, images, etc. However, still, people face difficulty in finding useful information or knowledge from those data. To address these challenges, recommendation systems come into the picture by providing useful content to the user based on users’ history and similarity among users. Content-based and collaborative filtering are two major building blocks of recommendation systems. Recommendation systems have been applied into numbers of a domain such as recommending movies, music, course, literature, items, people, links, location, healthcare, agriculture. In the agriculture domain, appropriate crops to cultivate and selecting applicable pesticides based on land quality and types of crops are interesting factors to consider for a country like India. Initially, we review different types of recommendation systems along with its application area. Subsequently, we explore various parameters to evaluate recommendation systems followed by open issues and research challenges. We further study the work carried out by existing researchers in the said domain. As part of our contribution through this research, we have selected the Agriculture domain and proposed our algorithm for recommending crops based on various parameters. As an outcome of our contribution, a crop is recommended to farmers based on his land. Also, the system recommends a list of lands for a given crop. Using statistical analysis, we achieve accuracy from 93% to 97%.