22. Which one of the following terms is NOT the advantages of paging in memory management?
(A) same page size
(B) preventing internal fragmentation
(C) preventing external fragmentation
(D) enabling memory protection
(E) enabling memory sharing

18. All 10.0.0.0/24 datagrams leaving local network have same single source IP address 138.76.29.7, different source port numbers. What technology does such a mechanism rely on? (A) NAT (B) NNTP (C) SMTP (D) ICMP (E) PPTP

19. Consider the following C code: After the code is executed, what is the output? (A) 25 (B) 35 (C) 45 (D) 55 (E) 65

20. Consider an infix expression: (A+B)*C+D*(E-F). What is its prefix expression? (A) **A++BC-DEF (B) ++A*BC-*DEF (C) +*+ABC*D-EF (D) *+ABC+*-DEF (E) +*+ABC*-DEF

21. Consider the following C code: What is the output of the above program? (A) 100 236 306 254 165 (B) 100 220 330 261 150 (C) 105 224 328 239 165 (D) 100 220 337 239 165 (E) Segmentation fault

23. Which one of the following protocols is implemented on top of SSL/TLS? (A) HTTP (B) HTTPS (C) POP3 (D) SMTP (E) TELNET

24. Refer to Table 3 and Table 4, please indicate the CORRECT SQL statement that can output as depicted in Table 5. (A) SELECT * FROM movies (B) SELECT TOP 1 movies.movie, actors.name FROM movies INNER JOIN actors ON movies.actor = actors.id ORDER BY actors.salary DESC (C) SELECT movies.movie, actors.name FROM movies INNER JOIN actors ON movies.actor = actors.id ORDER BY actors.salary (D) All of the above can output the same results, as shown in Table 5. (E) None of the above can output the same results, as shown in Table 5.

25. Consider a graph as follows. Starting from node D, what is the visiting sequence of all nodes by using depth-first-search. If you have multiple choices, just follow the alphabetical order. (A) D, B, A, C, E, F, G, H, I (B) D, F, G, I, H, E, B, C, A (C) D, F, G, I, H, E, C, A, B (D) D, B, A, C, E, H, I, G, F (E) D, F, G, I, H, E, C, B, A

26. Which one of the following has the best time complexity? (A) log n (B) n/2 (C) n (D) n log n (E) 2n

27. Machine learning can be divided into supervised learning and unsupervised learning. The two problems most commonly dealt with in supervised learning are regression and classification. When you predict the target to be a continuous numerical value, we call it regression； if the target is discrete, we call it classification. Which one of the following common algorithms is NOT a classification algorithm? (A) SVM (B) Linear (C) Decision trees (D) KNN (E) Naive Bayes

28. Which of the following terms is NOT included in 5 V’s of Big Data? (A) Volume (B) Variety (C) Value (D) Velocity (E) Volatility