1623087480

The k-means clustering algorithm is a foundational algorithm that every data scientist should know. It is popular because it is simple, fast, and efficient. It works by dividing all the points into a preselected number (*k*) of clusters based on the distance between the point and the center of each cluster. The original k-means algorithm is limited because it works only in the Euclidean space and results in suboptimal cluster assignments when the real clusters are unequal in size. Despite its shortcomings, k-means remains one of the most powerful tools for clustering and has been used in healthcare, natural language processing, and physical sciences.

Extensions of the k-means algorithms include smarter starting positions for its *k* centers, allowing variable cluster sizes, and including more distances than Euclidean distance. In this article, we will focus on methods like PAM, CLARA, and CLARANS, which incorporate distance measures beyond the Euclidean distance. These methods are yet to enjoy the fame of k-means because they are slower than k-means for large datasets without a comparable gain in optimality. However, as we will see in this article, researchers have developed newer versions of these algorithms that promise to provide better accuracy and speeds than k-means.

For anyone who needs a quick reminder, StatQuest has a great video on k-means clustering.

For this article, we will focus on where k-means fails. Vanilla k-means, as explained in the video, has several disadvantages:

- It is difficult to predict the correct number of centroids (
*k*) to partition the data. - The algorithm always divides the space into
*k*clusters, even when the partitions don’t make sense. - The initial positions of the
*k*centroids can affect the results significantly. - It does not work well when the expected clusters differ in size and density.
- Since it is a centroid-based approach, outliers in the data can drag the centroids to inaccurate centers.
- Since it is a hard clustering method, clusters cannot overlap.
- It is sensitive to the scale of the dimensions, and rescaling the data can change the results significantly.
- It uses the Euclidean distance to divide points. The Euclidean distance becomes ineffective in high dimensional spaces since all points tend to become uniformly distant from each other. Read a great explanation here.
- The centroid is an imaginary point in the dataset and may be meaningless.
- Categorical variables cannot be defined by a mean and should be described by their mode.

The above figure shows an example of k-means clustering of the mouse data set using k-means, where k-means performs poorly due to varying cluster sizes.

Instead of using the mean of the cluster to partition, the medoid, or the most centrally located data point in the cluster can be used to partition the data points; The medoid is the least dissimilar point to all points in the cluster. The medoid is also less sensitive to outliers in the data. These partitions can also use arbitrary distances instead of relying on the Euclidean distance. This is the crux of the clustering algorithm named Partition Around Medoids (PAM), and its extensions CLARA and CLARANS. Watch this video for a succinct explanation of the method.

In short, the following are the steps involved in the PAM method (reference):

The time complexity of the PAM algorithm is in the order of *O(k(n - k)*^{2}*)*, which makes it much slower than the k-means algorithm. Kaufman and Rousseeuw (1990) proposed an improvement that traded optimality for speed, named CLARA (Clustering For Large Applications). In CLARA, the main dataset is split into several smaller, randomly sampled subsets of the data. The PAM algorithm is applied to each subset to obtain the medoids for each set, and the set of medoids that give the best performance on the main dataset are kept. Dudoit and Fridlyand (2003) improve the CLARA workflow by combining the medoids from different samples by voting or bagging, which aims to reduce the variability that would come from applying CLARA.

Another variation named CLARANS (Clustering Large Applications based upon RANdomized Search) (Ng and Han 2002) works by combining sampling and searching on a graph. In this graph, each node represents a set of *k* medoids. Each node is connected to another node if the set of *k* medoids in each node differs by one. The graph can be traversed until a local minimum is reached, and that minimum provides the best estimate for the medoids of the dataset.

Schubert and Rousseeuw (2019) proposed a faster version of PAM, which can be extended to CLARA, by changing how the algorithm caches the distance values. They summarize it well here:

“This caching was enabled by changing the nesting order of the loops in the algorithm, showing once more how much seemingly minor-looking implementation details can matter (Kriegel et al., 2017). As a second improvement, we propose to find the best swap for each medoid and execute as many as possible in each iteration, which reduces the number of iterations needed for convergence without loss of quality, as demonstrated in the experiments, and as supported by theoretical considerations. In this article, we proposed a modification of the popular PAM algorithm that typically yields an O(k) fold speedup, by clever caching of partial results in order to avoid recomputation.”

In another variation, Yue et al. (2016) proposed a MapReduce framework for speeding up the calculations of the k-medoids algorithm and named it the K-Medoids++ algorithm.

More recently, Tiwari et al. (2020) cast the problem of choosing *k* medoids into a multi-arm bandit problem and solved it using the Upper Confidence Bound algorithm. This variation was faster than PAM and matched its accuracy.

#2020 dec tutorials #overviews #algorithms #clustering #explained

1621443060

This article provides an overview of core data science algorithms used in statistical data analysis, specifically k-means and k-medoids clustering.

Clustering is one of the major techniques used for statistical data analysis.

As the term suggests, “clustering” is defined as the process of gathering similar objects into different groups or distribution of datasets into subsets with a defined distance measure.

*K-means* clustering is touted as a foundational algorithm every data scientist ought to have in their toolbox. The popularity of the algorithm in the data science industry is due to its extraordinary features:

- Simplicity
- Speed
- Efficiency

#big data #big data analytics #k-means clustering #big data algorithms #k-means #data science algorithms

1618449987

For this week’s data science career interview, we got in touch with Dr Suman Sanyal, Associate Professor of Computer Science and Engineering at NIIT University. In this interview, Dr Sanyal shares his insights on how universities can contribute to this highly promising sector and what aspirants can do to build a successful data science career.

With industry-linkage, technology and research-driven seamless education, NIIT University has been recognised for addressing the growing demand for data science experts worldwide with its industry-ready courses. The university has recently introduced B.Tech in Data Science course, which aims to deploy data sets models to solve real-world problems. The programme provides industry-academic synergy for the students to establish careers in data science, artificial intelligence and machine learning.

“Students with skills that are aligned to new-age technology will be of huge value. The industry today wants young, ambitious students who have the know-how on how to get things done,” Sanyal said.

#careers # #data science aspirant #data science career #data science career intervie #data science education #data science education marke #data science jobs #niit university data science

1620466520

If you accumulate data on which you base your decision-making as an organization, you should probably think about your data architecture and possible best practices.

If you accumulate data on which you base your decision-making as an organization, you most probably need to think about your data architecture and consider possible best practices. Gaining a competitive edge, remaining customer-centric to the greatest extent possible, and streamlining processes to get on-the-button outcomes can all be traced back to an organization’s capacity to build a future-ready data architecture.

In what follows, we offer a short overview of the overarching capabilities of data architecture. These include user-centricity, elasticity, robustness, and the capacity to ensure the seamless flow of data at all times. Added to these are automation enablement, plus security and data governance considerations. These points from our checklist for what we perceive to be an anticipatory analytics ecosystem.

#big data #data science #big data analytics #data analysis #data architecture #data transformation #data platform #data strategy #cloud data platform #data acquisition

1617961560

The buzz around data science has sent many youngsters and professionals on an upskill/reskilling spree. Prof. Raghunathan Rengasamy, the acting head of Robert Bosch Centre for Data Science and AI, IIT Madras, believes data science knowledge will soon become a necessity.

IIT Madras has been one of India’s prestigious universities offering numerous courses in data science, machine learning, and artificial intelligence in partnership with many edtech startups. For this week’s data science career interview, Analytics India Magazine spoke to Prof. Rengasamy to understand his views on the data science education market.

With more than 15 years of experience, Prof. Rengasamy is currently heading RBCDSAI-IIT Madras and teaching at the department of chemical engineering. He has co-authored a series of review articles on condition monitoring and fault detection and diagnosis. He has also been the recipient of the Young Engineer Award for the year 2000 by the Indian National Academy of Engineering (INAE) for outstanding engineers under the age of 32.

Of late, Rengaswamy has been working on engineering applications of artificial intelligence and computational microfluidics. His research work has also led to the formation of a startup, SysEng LLC, in the US, funded through an NSF STTR grant.

#people #data science aspirants #data science course director interview #data science courses #data science education #data science education market #data science interview

1611381728

**Data Science** becomes an important part of today industry. It use for transforming business data into assets that help organizations improve revenue, seize business opportunities, improve customer experience, reduce costs, and more. Data science became the trending course to learn in the industries these days.

Its popularity has grown over the years, and companies have started implementing data science techniques to grow their business and increase customer satisfaction. In online Data science course you learn how Data Science deals with vast volumes of data using modern tools and techniques to find unseen patterns, derive meaningful information, and make business decisions.

**Advantages of Data Science**:- In today’s world, data is being generated at an alarming rate in all time lots of data is generated; from the users of social networking site, or from the calls that one makes, or the data which is being generated from different business. Because of that reason the huge amount of data the value of the field of Data Science has many advantages.

**Some Of The Advantages Are Mentioned Below**:-

**Multiple Job Options** :- Because of its high demand it provides large number of career opportunities in its various fields like Data Scientist, Data Analyst, Research Analyst, Business Analyst, Analytics Manager, Big Data Engineer, etc.

**Business benefits**: - By Data Science Online Course you learn how data science helps organizations knowing how and when their products sell well and that’s why the products are delivered always to the right place and right time. Faster and better decisions are taken by the organization to improve efficiency and earn higher profits.

**Highly Paid jobs and career opportunities**: - As Data Scientist continues working in that profile and the salaries of different position are grand. According to a Dice Salary Survey, the annual average salary of a Data Scientist $106,000 per year as we consider data.

**Hiring Benefits**:- If you have skills then don’t worry this comparatively easier to sort data and look for best of candidates for an organization. Big Data and data mining have made processing and selection of CVs, aptitude tests and games easier for the recruitment group.

**Also Read: How Data Science Programs Become The Reason Of Your Success**

**Disadvantages of Data Science**: - If there are pros then cons also so here we discuss both pros and cons which make you easy to choose Data Science Course without any doubts. Let’s check some of the disadvantages of Data Science:-

**Data Privacy**: - As we know Data is used to increase the productivity and the revenue of industry by making game-changing business decisions. But the information or the insights obtained from the data may be misused against any organization.

**Cost**:- The tools used for data science and analytics can cost tons to a corporation as a number of the tools are complex and need the people to undergo a knowledge Science training to use them. Also, it’s very difficult to pick the right tools consistent with the circumstances because their selection is predicated on the proper knowledge of the tools also as their accuracy in analyzing the info and extracting information.

#data science training in noida #data science training in delhi #data science online training #data science online course #data science course #data science training