1.0 What is correlation? In non mathematical terms, it is easy to understand correlation using the below picture. Do two things move together? Do they move independently of each other? Do they move opposite of each other? We can intuitively understand it. The number of hours put in to study is correlated to marks. Net caloric intake to weight is possibly high correlation. We can infer these because in this case inherent causation is clear.
0.0 The tangentially relevant cartoon 1.0 Relevance of text data The written word is all around us. In any organisation, large or small, there would always be a component of data which is textual. It could be reports from inspections, comments from customers, tweets regarding a product, search history of a user comments from employees email patterns of client logs generated by machines logs generated by logs generated by machines There is knowledge in text, but like Dilbert says, maybe it comes with strings attached.
1.0 What is distance? Basically, distance means a quantitative measure of how far apart two objects are in space. I am sure you know that. Spatial distance has been part of human life in perception since time eternal, and has been quantified and standardized across world over last few thousand years at least. Another way to think of this, particularly in terms of statistics, is dissimilarity. Closer means similar, and further means dissimilar.
1. Feature Selection : Part 2 This is part 2 of the blog on feature selection. In part 1, we discussed about what is feature selection, why do we need feature selection, different methods for feature selection and filtering based methods (https://r2dldocs.z6.web.core.windows.net/doc-repo/blog/feature-selection/). In this article we will continue with other methods Wrapper Methods Embedded Methods Hybrid Methods 2. Wrapper Methods Unlike basic filter methods, this class of methods use machine learning model for feature selection.
1. Introduction We all would have heard “Garbage In, Garbage Out”, in Machine Learning it can be read as “Noise In, Noise Out”. With growing data in all the fields it is important to understand the negative influence of noisy data on model’s accuracy and their demand for computational resources. Therefore, feature selection is an important step in any ML pipeline aimed at removing irrelevant, redundant and noisy features. Formally feature selection is the process of selecting a subset of relevant features for model development.
Bayesian Linear Regression using PyMC3 Introduction In this post we will talk about the motivation behind using Bayesian methods for machine learning models and we will look at a simple Bayesian linear regression model developed using the PyMC3 python library. How big a baby? Say you are working as a statistician specializing in public health. As part of your job, every month, you collect data on height and weight of babies who are less than 2 years old.
Author details Rohit Pruthi, Decision Scientist @ R2DL References AI for HR - Linked in learning course Networkx library Why data driven organization design We all know the collective power of the group. There are stories abound across the world on how teams accomplish goals, be it sports, corporate or even politics. Over time, we have inherently realized the power of the group. However, making effective teams has remained more of an art than a science.
We started working on a kaggle competition which deals with audio data. This notebook provides the learnings from audio analytics field. 1.0 Basics of Audio Analysis 1.1 What is sound? Audio signals are transverse pressure fluctuations (compressions and rarefactions of air pressure). When someone talks, it generates air pressure signals. The number of times the compression/rarefaction happens is known as the frequency of sound wave.
Delivery is the outcome phase of the Design Thinking methodology. This culminates with the tangible solution that is harnessed and refined based on the previous phases of discovery and design. Firstly, it’s not new Delivery is not a new terminology or concept that Design Thinking has coined. It has always been existence and practice across industries and domains. If delivery is a conventional methodology why does a contemporary approach refer to it or continue to refer to it?
Design phase is the second blog of this series and is the first step towards a tangible solution to any problem. This phase has three key activities (i) Ideation (ii) Prototyping and (iii) Validation that will incrementally move towards a robust solution. Design phase is the trickiest of the three as practitioners tend to fall trap for certain pitfalls or often overlap into other phases. So before getting to know more about the activities within the Design phase the top three cautions for anyone who is starting of this phase.
The very word Discovery does inflict a thought of finding something and this blog attempts to do nothing different rather within the boundaries of business problems irrespective their quantum. In order to have a better background of Discovery it is important to start with Design Thinking approach. A simple preface on Design Thinking Design Thinking bears quite a lot of definitions on the internet and it isn’t too difficult to make sense of them with a quick read through.
PowerApps provides a rapid low code development environment for building custom apps for business needs. It has services, connectors, and a scalable data service and app platform (Common Data Service) to allow simple integration and interaction with existing data. Power Apps enables the creation of web and mobile applications that run on all devices.People use apps for every area of their lives, and business should be no exception. Most out of the box solutions do not meet exact business needs or integrate well with other business programs.
Power Automate is a tool by Microsoft to enable users to create and automate process and actions that saves a lot of time and efforts for repetitive tasks. It helps you to create Automated, Instant, Scheduled, Business and UI Flows. Power automate allows users to integrate with other apps and services by means of connectors. Connector links 2 apps together so that data can be moved/flow between applications. Power Automate has 3 main components