The backup method that captures only the changes made since the last full backup is called an "Incremental Backup." It helps in conserving storage space and time by backing up only the data that has changed since the last backup, whether it's a full or incremental one.

In data warehousing, a cold backup is preferable to a hot backup when data needs to be backed up without disrupting the normal operations of the data warehouse. Unlike hot backups, cold backups can be taken when the system is offline, making them ideal for maintaining data integrity without interruptions.

Data profiling is a common tool or method used for monitoring the health and performance of a data warehouse. Data profiling helps in assessing data quality, identifying anomalies, and ensuring that data conforms to the expected standards. It is essential for maintaining data warehouse data integrity.

A federated data warehouse approach allows data warehouses in different regions to operate independently while also providing the capability to consolidate data when needed. It enables each regional data warehouse to maintain autonomy over its data, schema, and operations while still allowing global analysis and consolidation when required.

During the recovery of a data warehouse, the process of applying logs to restore the database to a consistent state is known as data loading. This process involves reapplying the transaction logs to recreate the state of the database at the time of the failure or recovery point.

In a three-tier data warehouse architecture, the middle tier typically contains the data transformation layer. This layer is responsible for ETL (Extract, Transform, Load) processes, data cleansing, and ensuring data consistency before it is presented to users.

This instability of parameter estimates is a typical symptom of multicollinearity. When predictors are highly correlated, it becomes hard for the model to determine the effect of each predictor independently, hence slight changes in data can lead to very different parameter estimates.

A Pearson's correlation coefficient value of 0 indicates no relationship between the two variables. Pearson's correlation measures the linear relationship between two datasets. A value of 0 suggests no linear relationship.

Continuous data can take any value within a range and can be subdivided infinitely.

Treating discrete data as continuous can lead to inaccurate conclusions due to incorrect statistical analyses. For example, it can affect the choice of statistical tests or machine learning models, leading to potential misinterpretation of the data.

Kendall's Tau is commonly used for ordinal data. It measures the ordinal association between two measured quantities. Like Spearman's correlation, it's based on ranks and is a suitable measure of association for ordinal data.

A "Low Standard Deviation" implies that the data values are "Close to the mean". In other words, most data points are close to the average data point.