VL 1.1 Business Analytics: Big Data & Statistics

Descriptive Statistics involves using machine learning techniques to identify future outcomes.

Prescriptive Statistics analyzes past performance and trends to determine future actions.

Big Data Analytics only refers to the high volume of data, not data types and structures.

Predictive Statistics focuses on identifying the likelihood of future outcomes based on historical data.

The main purpose of Big Data Analytics is to provide historical data insights only.

Decision Trees are an example of Descriptive Statistics techniques.

As data type becomes more disorderly, less data preparation or pre-processing time is required.

Semi-structured data can be easily addressed by using parsing, a function available in Excel.

Quasi-structured data is defined as having consistent and easily interpretable formats.

Parsing semi-structured data may lead to issues if the data is not well-organized.

Successful parsing of semi-structured data allows for proper processing of the data.

Quasi-structured data requires no additional effort for formatting and can be easily handled.

Cropital is a healthcare-focused crowdfunding platform that uses Big Data to assess the creditworthiness of potential borrowers.

Financial institutions purchase data from data collectors and aggregators to optimize pricing strategies.

Statista collects data on topics such as consumer behavior, industry trends, and market research.

Retailers purchase data from data aggregators to assess credit risk.

Marketing agencies purchase data from data aggregators to develop personalized financial products.

Data Users/Buyers are entities that collect data on consumer behavior, industry trends, and market research.

Data profiling involves examining tables, files, and database structure before the extraction process.

Failure during the ETL process is impossible if the data is considered 'Too Dirty'.

Incorrect relationships between tables are examples of data quality issues related to database structures.

Data quality issues related to having two or more files about a single subject, with mismatched unique identifiers, are common in the ordering process.

Intervention to address data quality issues includes reviewing the source details manually and making corrections as necessary.

Conducting data gathering is the first step in the Extract-Transform-Load (ETL) process.

A perceptron is a plane that maximizes the margin between two classes.

Support Vectors are equivalent to two lines at the edge of a plane.

The perceptron separates classes in a non-parametric way.

The perceptron is used to determine the class of any particular dot based on its position relative to the line.

Default values classified as 'yes' are represented by negative numbers relative to the perceptron.

Pre-processing of data involves plotting distance normalized values after normalization.