West Virginia University
Application of Neural Network-Support Vector Technique to Forecast U.S. Unemployment Rate
Abstract
dc:description.abstractThis research utilized five economic factors; 1) Consumer Price Index, 2) Return on Treasury Securities, 3) Total Nonfarm payroll, 4) Jobless Claims Filed, and 5) Stand & Poor 500 index to predict US unemployment rate. Historical time series data was obtained from the Economic Research web site of the Federal Reserve Bank of St. Louis and other finance web site.;Multiple Linear Regression, Back Propagation Algorithm, and Support Vector Regression techniques were utilized to predict US unemployment rate. Based on Mean Squared Error and adjusted R2 values, the Support Vector Regression technique provided superior results for the given dataset. Future US unemployment rate was predicted with an average absolute error value of 0.815, 0.13 and 0.07 using MLR, ANN and SVR, respectively.
Degree
thesis:*- Name thesis:degree_name
- MS
- Level thesis:degree_level
- Thesis
- Discipline thesis:degree_discipline
- Industrial and Managements Systems Engineering
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Ansari, Azadeh
- Contributors dc:contributor
-
- Rashpal S Ahluwalia
- Robert C Creese
- Feng Yang
Identifiers
dc:identifier.*- Identifier
- https://researchrepository.wvu.edu/etd/5112