We report the effectiveness of LU factorization (also called LU decomposition) as a technique for feature extraction along with naïve Bayes Classifier in recognizing handwritten Odia numerals (a regional language of north-eastern states of India derived from the Devanagari script). Experimental results show that LU factorization could be an alternative choice for feature extraction in pattern classification problems. Keywords: Handwritten Odia Numerals, LU Factorization, Naïve Bayes Classifier

Precious metals like gold are in high demand due to their usage as jewellery, storage of value and also as part of a diversified investment portfolio. When it comes to an investment option, it is obvious for the investors to know the expected returns in advance. Researchers apply a number of statistical and machine learning techniques to predict the future expected return of investment on gold. However, accuracy in prediction is still a challenging task due to the volatility in the market affecting parameters. The present study intended to analyse the effectiveness of a machine learning based hybrid model in predicting the future price of gold. Artificial Neural Network (ANN) based on Particle Swarm Optimization (ANN-PSO) model has been tested on monthly gold price in India for the period of January 2012 to June 2021. The experimental results of the present study investigated that the ANN-PSO model is capable of predicting the future gold price with high accuracy.