Question 1 (25%) I Answer the following questions with a True or False (15 marks). 1) In a broad sense, a knowledge-based system consists of a database and a knowledge base 2) A Prolog interpreter can process and respond directly to programs written in Prolog in an interactive approach 3) To convert numeric attributes to categorical ones, the normalization process must be applied to the numeric attributes OneR rules can be transformed into a linear regression model which can be used for the same purpose 5) A set of association rules is equivalent to a set of production rules for regression 6) Neural network classifiers are normally considered to be easy to interpret 7) A continuous attribute is a numeric attribute in a data mining/machine learning system 8) Disjoint training and test datasets are required to estimate the performance of classifers 9) K-means algorithms can automatically classify new instances into pre-defined categories 10) A common weakness of association rule mining is that it produces too many rules 11) Data sets represented by a matrix of instances versus attributes are suitable for machine learning and data mining applications 12) Visualization is a typical data preprocessing task 13) Machine learning methods can accept missing values 14) The Weka machine learning workbench was written in Prolog 15) Declarative languages in Artificial Intelligence are highly recursive Il. Multiple choice questions (1 mark each). Choose one answer per question. Wrong answers are penalized, i.e. one mark is deduced for each incorrect response. Indicate your choice (the choice letter) in Capital Letters on your answer sheet. DO NOT USE this sheet to provide your answers. USE YOUR SCRIPTBOOK only (total: 5 marks) 1) In a machine learning system, an instance (example) is characterised by a. A file containing all the data under study b A single PROLOG atom c. The values of the attributes d. Human expert rule written in Java