Question 1

What is Tableau Public?

Accepted Answer

Tableau Public is a type of social portal for discovering, developing, and publicly sharing data visualisations online. This platform is free, and with the world’s biggest collection of data visualisations, developing analytical skills is quite simple. With Tableau Public, it is possible to attain unlimited data inspiration and design a type of portfolio (company or private) online.

Question 2

Do I require coding and programming skills to learn Tableau?

Accepted Answer

One major benefit of Tableau is that programming and coding skills are not mandatory. Visual best practices and basic VizQL technology convey data and translate the drag-and-drop actions to data queries via an intuitive interface. The Tableau platform presents limitless data exploration and profound insights.

Question 3

Which method can improve the accuracy of a linear regression model?

Accepted Answer

One of the most widespread methods to enhance the accuracy of a linear regression model is “The Outlier Treatment.” This method is quite useful for boosting accuracy because the regression is quite sensitive to outliers. Hence, it becomes crucial to treat outliers with proper values.

Question 4

Why is linear regression called linear?

Accepted Answer

The linear regression graph shows that it fits a straight line, minimising the inconsistencies between the predicted and the original output values. The relationship between the variables is linear. Francis Galton first used the term ‘regression’ in his 1866 paper entitled ‘Regression towards mediocrity in hereditary stature’. He only utilised the word in the perspective of regression toward the mean. Subsequently, the term was used by others to indicate linearity, and therefore, linear regression is called linear.

Question 5

Why is linear regression analysis used?

Accepted Answer

Linear regression analysis helps predict a variable's value depending on another variable's value. The variable whose value you want to predict is the dependent variable. The variable used to indicate the value of another variable is known as the independent variable. Linear regression analysis helps to know which predictors in a model are statistically essential and which are not. Moreover, this analysis can provide a confidence interval for every regression coefficient it estimates.

Question 6

How does linear regression work?

Accepted Answer

Linear regression predicts a dependent variable value (b) depending on the given independent variable (a). It models the linear relationship between one or more variables. Every observation comprises two values. One is for the dependent variable, and another is for the independent variable. Linear regression works such that it allows the model to predict outputs for inputs it has never observed before.

Question 7

What does it means by Positive and Negative Linear Relationship?

Accepted Answer

A regression line can feature a Positive or Negative Linear Relationship. If the dependent variable progresses on the Y-axis and the independent variable progresses on the X-axis, it is called a Positive linear relationship. Conversely, if the dependent variable’s value reduces on the Y-axis and the independent variable’s value increases on the X-axis, it is called a Negative linear relationship.

Question 8

How is the Cost function useful in Linear Regression?

Accepted Answer

The Cost function optimises the regression coefficients and measures the performance of a linear regression model. It helps find the accuracy of the mapping function, which maps the input variable to the output variable. Moreover, it enables you to determine the optimal values for a0 and a1 that offers the best fit line for given data points. The alternate name of this mapping function is the Hypothesis function.

Question 9

What is multiple regression analysis?

Accepted Answer

Multiple regression analysis is a statistical method for examining the relationship between a single dependent variable and multiple independent variables. Its key objective is to use those independent variables whose values can forecast the value of the single dependent variable.

Question 10

Can OLS be called linear regression?

Accepted Answer

Yes, Ordinary least squares (OLS) is a linear least squares method for assessing the unknown parameters within a linear regression model. It is the technique used to determine the simple linear regression of a given data set. OLS estimates the relationship between the variables by minimising the sum of the squares in the variance between the observed and predicted values of the dependent variable aligned as a straight line.

Question 11

How do linear regression and logistic regression differ from each other?

Accepted Answer

Linear Regression deals with regression problems, whereas Logistic regression deals with classification problems. Linear regression offers a continuous output, while Logistic regression offers discrete output. Linear Regression aims to find the best-fitted line, but Logistic regression fits the line values to the sigmoid curve. The method to calculate loss function in linear regression is mean square error but its maximum likelihood estimation in the logistic regression.

Question 12

When do you get the negative value of the linear regression coefficient?

Accepted Answer

You get a negative value of the linear regression coefficient when the value of the independent variable increases with the decrease in the value of the dependent variable. The negative coefficient value indicates how much the mean of the dependent variable differs when there is a one-unit change in the independent variable. In this evaluation, the values of other variables stay constant.

Question 13

When does overfitting occur in linear regression?

Accepted Answer

In linear regression, overfitting takes place when the model is very complex. Usually, this situation arises when there are more parameters than the number of observations. A linear regression model with overfitting will not perfectly generalise to new data. So, it will perform efficiently on training data but poorly on test data. Factors responsible for overfitting are –(i) outliers in the train data and (ii) train and test data belonging to different distributions.

Question 14

What is Curve Fitting with Linear Regression?

Accepted Answer

When using a linear regression model, the widespread way to fit curves for the data is to incorporate polynomial terms like cubed or squared predictors. Commonly, you need to select the model order based on the number of bends you require in your line. Every increment in the exponent generates one more bend into the curved fitted line.

Question 15

Why is linear regression not appropriate for time series?

Accepted Answer

One of the key assumptions of linear regression is that the residues aren’t correlated. This is often not the case with the time series data. In case there are autocorrelated residues, linear regression cannot capture all the trends within the data. Therefore, linear regression is generally not used for time series.

Question 16

Do outliers influence regression parameters?

Accepted Answer

Including outliers and influential cases can significantly alter the magnitude of regression coefficients. They can also change the coefficient signs, i.e. from negative to positive or vice versa. Their empirical results can be erroneous when abnormal observations are ignored, specifically concerning dependent variables.

Linear Regression Online Courses

Linear Regression Course Overview

Best Data Science Courses

Data Science (0)

upGrad Learner Support

Disclaimer