R Programming Interview Questions

R is a programming language and a software environment meant for statistical analysis and creating graphs. It is used by analysts, statisticians and data scientists for various purposes. R uses data objects for data calculations and it is an alternative to conventional statistical packages such as SAS, SPSS, etc. A lot of companies are incorporating R into their business models to proliferate their revenue. There is a huge career prospect in R such as data scientist, R programmer, Analyst consultant, etc.

The functions that R provides are

• Mean- it is calculated by taking the sum of the values and dividing it by a number of values. The function used is mean().
• Median- it is the middle most value in the data series. The function used in R programming is median().

Other functions of R include Regression, GLM, mixed-effects, distribution, GAM, non-linear, etc.

Data structure is a form of organizing and storing data. It is imperative to have a strong understanding of various data types and data structures in order to make the best use of R languages. R programming supports five types of data structures namely vector, matrix, list, data frame and factor.

• Vector– This data structure contains an integer, double, complex, etc. It is a sequence of same data elements and c() function is used to create a vector in R programming.
• Matrix- it is a two-dimensional data structure and is used to bind vectors from the same length. All the elements in the matrix have to be of the same type and it is created using a matrix() function. The value of row can be defined using nrow and the value of the column can be defined using ncol.
• List- list includes data of different types like numbers, strings, vectors, etc. It is somewhat like a vector but it contains mixed elements. A list is created using ().
• Data frame- it is a special list where each element is of the same length. A data frame has the features of both, matrices and lists. It is more generic than the matrix as different columns have different data types. It is crated using frame() function.
• Factors-it is created using factor() function and is used to store predefined and categorical data.

The following steps are followed to build and evaluate a linear regression model in R

• The first step is to divide the data into train and test sets. This step is crucial because you have to build a model on the train and evaluate its performance on the test set. This can be done with sample.split() function from the package.
• The second step is to build the model on the train set. The function used to build the model is Im() function.
• Once you’ve built the model, you can predict the values on the data set with the help of the predict() function.
• The last step in the linear regression model is to find out the RMSE. The lower the value of RMSE, better is the prediction.

R packages are the collection of R functions and sample data that are stored under a directory name called library. Initially, R adds a set of packages during installation. However, new packages are added as and when required for specific purposes. The different packages available in R are

• MICE- the MICE package deals with missing data. It creates replacement values for the missing data. There are two types of missing data namely MCAR and MNAR. In this package, the mice() function looks after the imputing process.
• Amelia- this package is used for multiple imputations of missing data. it also produces multiple output datasets for analysis. To use this package, you can either invoke ameliagui() command or run Amelia function on the data.
• Mi– the Mi package provides functions for data manipulation and imputes missing values. This package has several features that allow the users to get into the imputation process and gauge the reasonableness of the resulting model.

The functions in dplyr package are-

• Filter- Filter() allows you to select a subset of rows in a data frame. the first argument is the tibble and the second argument is the variables within that data frame. it selects the rows where expression is true.
• Arrange- Arrange() recorders the rows on the basis of data frames or a set of column names.  Desc() function is used to arrange columns in descending orders.
• Mutate- it is used to add new variables to the data. it is also used to add new columns that are functions of existing columns. Dplyr::mutate is used to refer to the newly created column.
• Select()- this function is used to zoom in on a useful subset that works on numeric values. With select(), you can use functions like ends_with(), matches(), starts_with(), etc.

The packages used for mining in R are-

• Data.table- supports fast reading of large files.
• Arules- used for rule learning.
• Tm- used to perform text mining.
• Forecast- provides functions for time series analysis.  

Clustering refers to the group of objects that belongs to the same class. It is a process to make a group of abstract objects into the class of similar objects. Clustering is required in data analysis due to the following reasons-

• Scalability- clustering is required to deal with large databases.
• Interpretability- the result of clustering should be comprehensive and usable.
• Dimensionality- the clustering algorithm is used to handle high-dimensional space.
• Deal with noisy data- Databases contains erroneous data. Algorithms that are sensitive to such data may deliver poor results.

• KMEANS clustering in this method, objects are classified as belonging to K-groups. It is also known as partitioning method. The result of this method is K clusters and in each cluster, there may be a centroid. This method is popular for cluster analysis in data mining. The K-means clustering algorithm is used to find groups which have not been labelled in the data. This method is used to find groups in the data, with the number of groups which are represented by K.
• Hierarchical clusteringIt is a method of cluster analysis which aims to build the hierarchy of clusters. This method has two approaches namely divisive approach and agglomerative approach. In agglomerative approach, each object forms a separate group and keeps on merging the groups that are close to one another. It is also known as the bottom-up approach. In a divisive approach, we start with all the objects in the same cluster. The cluster is split into smaller clusters. It is also known as a top-down approach.

Rattle gives statistical and visual summaries of data and is a popular GUI for data mining. It transforms data so it can be easily modelled and builds a supervised and unsupervised ML model from the data. It also gives the graphical presentation of the models. Rattle is also used as a teaching facility to learn R languages. The features of Rattle package include clustering, modelling, evaluation, statistical test, etc.

In R, a white noise model is a basic time series model which is also the basis for more elaborated and defined models. To stimulate the data from a variety of tie series model, Arima.sim() function is used. The white noise model has a fixed constant mean, fixed constant variance and no correlation over time.

In R programming, random walk model is an example of the non-stationary model. A random walk has no fixed mean or variance. It also has a strong dependence over time. There are two types of random walks namely random walk without drift and random walk with drift.

There are several ways to import data in R. You can use R commander to import data in R.

• Excel file- If the sample data is in excel format, function read.xls is used from the data package. It returns a data frame. Alternatively, loadWorkbook can also be used to read the entire workbook.
• Minitab file- if the data file is in Minitab format, it can be opened using the function read.mtb. It returns a list of components in the Minitab worksheet.
• SPSS file- the data files in SPSS formats can be opened using the function read.spss. It returns a list of components.
• CSV file- Each cell inside the CSV data file is separated by a special character such as a comma.

under the Principal Component Analysis, the data is transformed into a new space. The first principal component takes the maximum amount of variance from the original data. The second principal component captures the amount of variability left. This is true for each component element and they are all uncorrelated. In R programming, Principal Component Analysis can be done using the function prcomp().

The following are the few differences between Python and R language

Library()- If the desired package cannot be loaded, this function will display an error message. It loads the package whether it is already loaded or not.

Require()- When a particular package is not found, it gives warning messages. Require() is used inside a function. It checks whether it is loaded or not and loads if it is not loaded.

R language is currently the most sought-after programming languages. It offers several benefits to the users

• R is a comprehensive language used for manipulation and managing of data.
• R has good graphical capabilities.
• It is free and open source software.
• There are no licence restrictions for R.
• R runs on many operating systems and hardware.
• R supports all statistical tests and models.

In R, the following sorting algorithms are available

• Bubble sort
• Selection sort
• Merge sort
• Quick sort
• Bucket sort

The following are the programming features of R

• There are packages in R. these R packages are useful in collecting functions into a single set.
• R programming includes database input, data export, variable lables, etc.
• R is an interpreted language and support matrices.
• R supports object-oriented programming and procedural programming. Object-oriented programming consists of classes, objects whereas procedural programming includes records, procedures.

The scope of R as a programming language is high and it has varied applications in various verticals. The important applications are

• R is used as an important tool in finance. It is used by several data analysts and research programmers.
• R deals with a lot of statistics. It is considered the best fir data science. R also provides an environment for statistical computing and design.
• R is also used for data importing and cleaning.