Author : Stanley ziweritin, iduma aka ibiam, taiwo adisa oyeniran, godwin epiahe oko
Keyword : Artificial intelligence, decision tree, k-nearest neighbor, machine learning
Subject : Science and technology
Article Type : Original article (research)
Article QR Code
Article File : Full Text PDF
Abstract : Machine learning is one of the fast growing areas of interest in artificial intelligence adopted by professional in every spheres of life that uses algorithms with data to sytematically learn patterns and improve from experience. The increasing competitive and robust predicting methods of machine learning are becoming more interesting and popular. This is valuable to investors, surveyors and valuers against manually computed payment table values that depends on emperical results. There are tedious and rigorous processes in valuation practice that involves some aspects of financial analysis in computations for the one pound table values. The aim is to build K-nearest neigbr and decision tree model to predict the nemeric values in amount of one pound table at a give rate of interest and period of years.This model is useful to investors, accountants, data professionals, surveyors and valuers interested in financial analysis and its applications. A cross validation test was carried out with predicted R-squared test to detect overfitting and generalize model performance on testing dataset. We introduced noisy data with smoothing curve expeoneintial function to overcome the risk of overfitting in predicting target varaible. The K-nearest neighbor and decision tree techniques were trained, tested and resulted into 95.76% and 99.86% respectively.
Article by : stanley ziweritin
Article add date : 2021-07-31
How to cite : Stanley ziweritin, iduma aka ibiam, taiwo adisa oyeniran, godwin epiahe oko. (2021-July-31). Knn and decision tree model to predict values in amount of one pound table. retrieved from https://www.openacessjournal.com/abstract/793