Evaluation of Modified K-Means Clustering Algorithm in Crop Prediction

An agricultural sector is in need for well-organized system to predict and improve the crop over the world. The complexity of predicting the best crops is high due to unavailability of proper knowledge discovery in crop knowledge-based which affects the quality of prediction. In data mining, cluster...

Celý popis

Uloženo v:
Podrobná bibliografie
Vydáno v:International journal of advanced computer research Ročník 4; číslo 3; s. 799
Hlavní autoři: Narkhede, Utkarsha P, Adhiya, K P
Médium: Journal Article
Jazyk:angličtina
Vydáno: Bhopal Accent Social and Welfare Society 01.09.2014
Témata:
Algorithms
Cluster analysis
Clustering
Crops
Data mining
Knowledge base
Maximization
Tasks
ISSN:2249-7277, 2277-7970
On-line přístup:Získat plný text
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Popis
Shrnutí:An agricultural sector is in need for well-organized system to predict and improve the crop over the world. The complexity of predicting the best crops is high due to unavailability of proper knowledge discovery in crop knowledge-based which affects the quality of prediction. In data mining, clustering is a crucial step in mining useful information. The clustering techniques such as k-Means, Expectation Maximization, Hierarchical Micro Clustering, Constrained k-Means, SWK k-Means, k-Means++, improved rough k-Means which make this task complicated due to problems like random selection of initial cluster center and decision of number of clusters. This works demonstrates an evaluation of modified k-Means clustering algorithm in crop prediction. The results and evaluation show comparison of modified k-Means over k-Means and k-Means++ clustering algorithm and modified k-Means has achieved the maximum number of high quality clusters, correct prediction of crop and maximum accuracy count.
Bibliografie:SourceType-Scholarly Journals-1
ObjectType-Feature-1
content type line 14
ObjectType-Article-1
ObjectType-Feature-2
content type line 23
ISSN:2249-7277
2277-7970