Tag Archives: Autocorrelation Function

A Statistical Approach to Study Pitch Variation for Prediction of Voice Disorder (Published)

The human voice is a magical tool to communicate, express emotions, to create wonderful music through singing. Everyone has a distinct voice, different from all others; almost like a fingerprint, one’s voice is unique and can act as an identifier. Unfortunately we never give much attention to the voice problem that can limit our ability to communicate and to complete daily activities. Voice disorders have a higher rate of occurrence among those who are in professions that require speaking, like teachers, aerobatic instructors, lawyers, social workers etc. Symptoms of a voice disorder range from hoarseness or a chronic dry, scratchy throat, a pitch/tone that is not pleasing, limitations in the ability to speak clearly, or periods of voice loss. Voice disorder may be attributed to the abnormality in the structure and/or function of the larynx (vocal folds) as well as the abnormalities of its different components, making each voice different; namely, pitch, tone, and rate. In this paper a statistical approach is investigated on pitch quality to make early prediction of voice disorder for professional talkers.

Keywords: Autocorrelation Function, Electroglottography, Kymographic parameters, Phonation, Spectral density, Vocal folds

Forecasting of Exchange Rate between Naira and US Dollar Using Time Domain Model (Published)

Most time series analysts have used different technical and fundamental approach in modeling and to forecast exchange rate in both develop and developing countries, whereas the forecast result varies base on the approach used or applied. In these view, a time domain model (fundamental approach) makes the use of Box Jenkins approach was applied to a developing country like Nigeria to forecast the naira/dollar exchange rate for the period January 1994 to December 2011 using ARIMA model. The result reveals that there is an upward trend and the 2nd difference of the series was stationary, meaning that the series was I (2). Base on the selection criteria AIC and BIC, the best model that explains the series was found to be ARIMA (1, 2, 1). The diagnosis on such model was confirmed, the error was white noise, presence of no serial correlation and a forecast for period of 12 months terms was made which indicates that the naira will continue to depreciate with these forecasted time period.

Keywords: AIC, Auto Regressive Integrated Moving Average, Autocorrelation Function, BIC, Exchange Rate, Partial Autocorrelation Function