Unit 5: Introduction to Big Data and Data Analytics

In data analysis, a feature is a measurable attribute of a record, object, or example, such as a student's age, marks, or attendance. A dataset column can serve as a feature when it provides useful input information for identifying patterns or making predictions. The other options describe a graph, file format, or analysis result rather than an input attribute.

An outlier is a data value that is unusually large, small, or otherwise far from the general pattern of most values in a dataset. The exact method for identifying one can depend on the context and statistical method used. Option B is the mode, option C is the mean, and option D is the median.

AI systems may use names, contact details, academic records, or learning-related information. Privacy protects such data from unauthorized access, misuse, and disclosure without permission. Faster processing is a performance issue, not the main purpose of privacy. Privacy also supports informed consent and gives individuals greater control over their personal information.

Deployment is the stage in an AI project when a developed and evaluated model is integrated into an application, service, or operational process so that intended users or systems can use it. Deployment may also require monitoring, maintenance, updates, and safeguards after release. It is not the same as defining the problem, deleting training data, or rejecting an untested model. Thus, option A gives the correct meaning.

Correlation indicates the direction and strength of association between two variables. A positive correlation means that the variables tend to increase or decrease together, while a negative correlation means that one tends to rise as the other falls. A value near zero suggests little linear association. Option A is correct; correlation alone does not prove causation, and it is not itself the mean or spread.

The arithmetic mean represents the average of a numerical data set. It is calculated by adding every value and dividing the total by the number of values: mean = sum of values ÷ number of values. For example, the mean of 2, 4, and 6 is 12 ÷ 3 = 4. Therefore, option A is correct. The smallest and largest values are extrema, while the ordered middle value is the median.

The median is the middle value of an ordered dataset and is less affected by extreme values than the mean. It can therefore provide a more representative measure of central tendency when very large or very small outliers are present, although the best measure also depends on the dataset and purpose.

The mode is the value or category with the greatest frequency in a data set. For example, in 2, 3, 3, 4, the mode is 3 because it appears twice, more often than the other values. Option A is correct. A data set may have more than one mode or no unique mode. The least frequent value, largest value, and total sum describe different properties.

Descriptive statistics organize, summarize, and present the main features of collected data. Measures such as the mean, median, mode, range, and standard deviation, along with tables and charts, are commonly used. Forecasting future outcomes is generally associated with predictive analysis rather than descriptive statistics.

The spread of data indicates how widely the values vary around a central value, such as the mean or median. It therefore helps describe variability and distinguish datasets that may have similar averages but different levels of consistency. Spread does not assume that all values are equal.

Artificial intelligence can analyse transaction data to identify patterns that differ from normal behaviour, such as a payment of an unusual amount, at an unusual time, or from an unusual location. Such transactions can be flagged for review. Therefore, option A is correct. Unlike option B, AI does not approve every transaction without checking it.

Anomaly detection identifies data points or events that deviate from an expected or normal pattern. For example, an unusually large bank transaction may indicate possible fraud. Therefore, option A is correct; option B does not distinguish unusual events from normal ones. Exam tip: remember that an anomaly is something abnormal or different from the usual pattern.

When points in a scatter plot lie approximately along a straight line, they can indicate a linear relationship or correlation between two variables. An upward pattern may show positive correlation, while a downward pattern may show negative correlation. Option C is incorrect because correlation shows association, not necessarily causation.

Anonymization removes or changes names, identification numbers, or other identifying details so that a person cannot be directly identified while the data is used. Its main purpose is therefore to protect privacy and reduce the risk of revealing identity. It does not increase data accuracy or permanently delete all data. Exam tip: Associate anonymization with hiding identity and protecting data privacy.

Anomaly detection looks for unusual deviations from normal transaction patterns. A fraudulent transaction may show an unusual amount, time, location, or purchase pattern, so it can be flagged for further review. Option B is incorrect because even legitimate transactions are not exactly identical; anomaly detection aims to distinguish normal variation from suspicious deviations.