Sankar Vema - Blog Space

Over last two decades we have seen enterprise application landscape changing very rapidly with rapidly evolving technology stacks and changing industry dynamics. It is time for another paradigm shift in terms of how we conceptualize, design, develop, test and maintain our applications to meet volatile business requirements. Some of the prime features of this evolving paradigm Flexible business user centric infrastructure (than IT centric) IT as a service model with emphasis on user self service. Flexible development and deployment infrastructure which is readily available over cloud (no setup time, no high budgets and initial spending) Develop application using modern development languages which help improve the developer Usage of more dynamic meta programming languages Reusable assets, code generators, ... Build once and run anywhere across platforms and devices End-to-End application lifecycle integration through ALM and DevOps Improve communication and collaborat...

Further to my previous post on the webinars , session recordings now available online Jan 25th Cloud Computing Strategies for Enterprise You Tube Techgig Feb 2nd Cloud Implementation Scenarios for Enterprise You Tube Techgig

We are doing webinar series on Cloud Computing Strategies this month. You can join at the below links 25 th  Jan session:  Cloud Computing Strategies for Enterprise 2 nd  Feb session:  Cloud Implementation Scenarios for Enterprise Recordings of above webinars are available online shortly.

Last 8 months were been really hectic with the delivery commitments which kept me away from my blogs. Taking sometime, thinking of reviving my blog postings which were offline for quite some time due to expired membership with my hosting provider. Till I could make some proper arrangements on that I am thinking of posting all my old blogs here .

Last week, during the lunch meeting, I was talking to Greg (IT director of a manufacturing major) on different IT management strategies in the manufacturing sector. Slowly Greg started explaining his challenges in  meeting the business expectation of maintaining optimal IT service while balancing the current outdated technical landscape. One of his biggest concern was on the increasing cost of total ownership which is getting compounded due to rapidly changing IT technologies. Frankly, this is not just Greg's challenge but this is the challenge of every IT manager.     Discussing this topic in the evening with my colleague we decided to launch a blog on this areas so that we can share our experiences and ideas with like other in this industry. This gave birth to IT Next Gen a blog on the future of IT and different strategies for the technical managers which will help them keeping themselves flexible to the rapid changes while meeting the business expect...

As discussed in my previous blog, Imbalanced data poses serious challenges in Machine Learning .  One of approach to combat this imbalance is data is to alter the training set in such a way as to create a more balanced class distribution so that the resulting sampled data set can be used with traditional data-mining algorithms. This can be achieved through...  Under-sample where the size of the majority class is reduced using different techniques like reducing redundancy, removing boundary candidates etc., Over-sample where the size of the minority class is increased by adding more candidates which can augment the data set. Hybrid approach where a combination of both oversampling of minority class and under sampling of majority class is attempted. Each of these techniques discussed below Random Over Sampling In random over-sampling, the minority class instances are duplicated in the data set until a more balanced distribution is reached. As a illustration, consider...

For many real world machine learning problem we see an imbalance in the data where one class under represented in relative to others. This leads to mis-classification of elements between classes. The cost of mis-classification is often unknown at learning time and can be far too high. We often see this type of imbalanced classification scenarios in fraud/intrusion detection, medical diagnosis/monitoring, bio-informatics, text categorization and et al. To better understand the problem, consider the “Mammography Data Set,” a collection of images acquired from a series of mammography examinations performed on a set of distinct patients. For such a data set, the natural classes that arise are “Positive” or “Negative” for an image representative of a “cancerous” or “healthy” patient, respectively. From experience, one would expect the number of noncancerous patients to exceed greatly the number of cancerous patients; indeed, this data set contains 10,923 “Negative” (majority class) and...