Match the following: No. List-I Code List-II (i) OLAP (a) Regression (ii) OLTP…

Concept — this item tests two independent classification axes, and each axis must be applied on its own before any pairing is read off.

• Processing style of a database workload: OLTP (Online Transaction Processing) denotes the day-to-day operational workload — many short, concurrent, ACID read/write transactions over current, normalised data, which is what a general-purpose relational database system is built to serve. OLAP (Online Analytical Processing) denotes the analytical workload — few long, read-mostly, multidimensional aggregate queries over historical, integrated, subject-oriented data, which is what a data warehouse is built to serve.

• Data-mining task by the type of value a model predicts: predicting a discrete class label is classification; predicting a continuous numeric quantity is regression.

Application — apply each axis to the four left-hand entries in turn.

1. OLAP is the analytical processing style, so its storage counterpart is the Data Warehouse, whose integrated historical subject-oriented design exists precisely to answer multidimensional aggregate queries.

2. OLTP is the transactional processing style, so its storage counterpart is the RDBMS, whose normalised schemas, indexes and ACID transactions exist precisely to serve high-volume short operational transactions.

3. A Decision Tree recursively partitions the attribute space and attaches a discrete class label to each leaf, so the task it names on the second axis is Classification.

4. A Neural Network with a continuous-valued output unit fits a numeric response surface, so the task it names on the second axis is Regression.

Cross-check — the two axes do not overlap, so the pairing is forced rather than chosen. Once the two database-workload entries have consumed Data Warehouse and RDBMS, only Regression and Classification remain for the two learning models; the discrete-leaf output of the Decision Tree fixes it to Classification, which leaves Regression for the Neural Network. Reading the codes in order gives (b), (c), (d), (a).

Note — in practice both learning models are general purpose: regression trees predict continuous values and neural networks with a softmax output classify. This item uses the canonical textbook association, confirmed by the official examination key, in which the tree is the standard classification learner and the network with a linear output is the standard regression learner, and each code may be used only once, so the mapping is unambiguous.

Answer — (i)-(b), (ii)-(c), (iii)-(d), (iv)-(a).