Predictive Analytics with TensorFlow 2.1: Using Statistics in Predictive Modeling

The video tutorial covers the use of statistics in predictive modeling, explaining the role of statistical models, both parametric and non-parametric. It discusses the concepts of population and sampling, including the central limit theorem, and delves into descriptive statistics like mean, median, mode, variance, and standard deviation. The tutorial also explores covariance and correlation, interquartile range, hypothesis testing, and chi-square tests, providing practical examples and code snippets using Python libraries like NumPy and SciPy.

What is the primary purpose of using statistics in predictive modeling?

• To create visualizations

• To enhance security

• To predict outcomes

• To store data

Which type of predictive model has a fixed number of parameters regardless of data size?

• Non-parametric

• Parametric

• Descriptive

• Inferential

What does random sampling ensure in a population?

• Equal probability of selection

• Higher accuracy

• Faster processing

• Larger sample size

According to the central limit theorem, what does a sufficiently large sample size approximate?

• The mean of the population

• The range of the population

• The mode of the population

• The median of the population

What is the square root of variance known as?

• Mean

• Median

• Standard deviation

• Mode

What does covariance measure?

• The average of a data set

• The central tendency of a data set

• The spread of a data set

• The extent to which two variables change together

What is the range of correlation values?

• -∞ to ∞

• -1 to 1

• 0 to 1

• 0 to ∞

What is the null hypothesis in hypothesis testing?

• A hypothesis that is always false

• A hypothesis we want to prove

• A generally accepted fact

• A hypothesis with no evidence

Which test is used to determine the relationship between categorical variables?

• ANOVA

• Chi-square test

• Z-test

• T-test

What does a p-value less than 0.05 typically indicate in a chi-square test?

• Variables are dependent

• Variables are independent

• Sample size is too small

• Test is invalid