Application of Labelled K Means Clustering for GIS Contract Automation

Muhammad Shaheen

Abstract:

K Means Clustering is an unsupervised classification technique which is suitable for the dataset which do not have any class labels. The absence of labels constrained its application on different text data sets. Labelled K Means clustering generate labels for the clusters obtained from K Means Clustering which makes this technique more decisive for text data sets. A novel application of Labeled K Means Clustering for automation of technical part of GIS contracts is given in this paper. Geographic Information System (GIS) contract which is signed between GIS service provider and a client requires an efficient tracking system during all its vital phases (development, operation and maintenance). At present, the tracking is done in a non-discrete manner through manual inspection. Manual tracking is inevitable because (1) no indicators have so far been developed for evaluation by an automatic tracking system, (2) No automated system exists to evaluate the performance of GIS service provider and the client on the basis of performance indicators and (3) there is no centralized mechanism for penalizing negligence of either party. This paper proposes (1). a method to regulate the technical part of the GIS contract by suggesting a simple and wizard-based Graphical user interface. (2). Conversion of existing manually prepared contracts into electronic contracts through lexeme-based congregation which is done through labelled K Means Clustering. These converted clusters are then stored into centralized database. Back Propagation Neural Network (BPNN) is used to train the system on performance indicators defined for compliance by both contracting parties.

research from:

• Foundation University - Faculty of Engineering and IT

Year:

2019

Type of Publication:

Article

Volume:

38

Number:

1

Pages:

15-22

Month:

6

DOI:

10.25211/JEAS.V38I1.2900