Comparison of machine learning algorithms over prediction of Titanic database

Abstract

With the popularize of artificial intelligent, algorithms such as machine learning algorithm takes an important part in many fields. In terms of data prediction, it is important to know which machine learning algorithm fits the specific dataset the most. Comparison between machine learning algorithms helps us figure out the advantage and disadvantage of them and thus save time, and money during the working process. In this paper, this paper will be comparing K-nearest neighbours algorithm, Random Forest algorithm, and Support vector machine algorithm mostly focusing on their performance over the survival prediction of Titanic dataset. It could be seen that algorithm with more adjustable hyper parameters could produce more accurate result. However, finding the suitable hyper parameters is time consuming. In the future, this paper may try to explore ensembled algorithms for better result.