<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Bhavesh Bhatt on Medium]]></title>
        <description><![CDATA[Stories by Bhavesh Bhatt on Medium]]></description>
        <link>https://medium.com/@bhattbhavesh91?source=rss-a750aa010abc------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*FjjnX5IKCx4UGWrt66Slfw.jpeg</url>
            <title>Stories by Bhavesh Bhatt on Medium</title>
            <link>https://medium.com/@bhattbhavesh91?source=rss-a750aa010abc------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Wed, 18 Nov 2026 02:12:35 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@bhattbhavesh91/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[[ML Story] How I cleared the Google Cloud Machine Learning Certification in 2023]]></title>
            <link>https://medium.com/google-developer-experts/ml-story-how-i-cleared-the-google-cloud-machine-learning-certification-in-2023-15f3bdef61c1?source=rss-a750aa010abc------2</link>
            <guid isPermaLink="false">https://medium.com/p/15f3bdef61c1</guid>
            <category><![CDATA[certification]]></category>
            <category><![CDATA[google]]></category>
            <category><![CDATA[google-cloud-platform]]></category>
            <category><![CDATA[machine-learning]]></category>
            <category><![CDATA[google-developer-expert]]></category>
            <dc:creator><![CDATA[Bhavesh Bhatt]]></dc:creator>
            <pubDate>Fri, 13 Oct 2023 17:21:15 GMT</pubDate>
            <atom:updated>2023-10-13T17:21:15.777Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*N0iXM7sr1qpKZvQOoRCqQQ.png" /></figure><p>Google Cloud Machine Learning Engineer Exam is a certification offered by Google Cloud for individuals who want to demonstrate their expertise in designing, building, and deploying machine learning models on the Google Cloud Platform (GCP). This certification is intended for professionals who work with machine learning technologies and want to validate their skills in the context of GCP. The Google Cloud Machine Learning Engineer certification is one of the most challenging cloud certifications available. It covers a wide range of topics, including machine learning fundamentals, Google Cloud Platform services, and MLOps practices.</p><p>I recently cleared the certification exam, and in this article, I will share my preparation strategy and tips to help you succeed.</p><p><strong>Prerequisites:<br></strong>Before diving into the preparation, it’s essential to understand the basics of the Google Cloud Machine Learning Certification. This certification is designed for professionals who want to showcase their expertise in using Google Cloud services to design, build, and deploy machine learning models. It covers a wide range of topics, including data preparation, model building, deployment, and MLOps practices. Before you attempt the exam, it is recommended that you have the following:</p><ul><li>3+ years of hands-on experience with GCP products and solutions.</li><li>Strong understanding of machine learning fundamentals, such as linear regression, logistic regression, decision trees, and random forests.</li><li>Experience with Python programming language.</li></ul><p><strong>Preparation Strategy :<br></strong>My preparation strategy for the exam was as follows:</p><ol><li><strong>Review the exam guide </strong>: The exam guide provides a detailed overview of the topics covered in the exam. Make sure to thoroughly review the guide and identify any areas where you need to focus your studies.</li><li><strong>Research &amp; Resources </strong>: I searched the internet for resources, including official Google Cloud documentation, online courses, and practice exams.</li><li><strong>Learning Path</strong> : There are many excellent training courses available online. I referred to the book <strong>Journey to Become a Google Cloud Machine Learning Engineer: Build the mind and hand of a Google Certified ML professional </strong>and I watched a lot of YouTube videos<strong>.</strong></li><li><strong>Hands-On Practice </strong>: Theory alone is not enough. I set aside dedicated time for hands-on practice with GCP services. I also recommend completing the <strong>Qwiklabs</strong> for the Professional Machine Learning Engineer certification.</li><li><strong>Take practice tests</strong> : There are many practice tests available online and in the <strong>Google Cloud Skills Boost platform</strong>. Taking practice tests will help you identify your strengths and weaknesses, and track your progress over time.</li><li><strong>Build your own machine learning projects</strong> : This is the best way to learn by doing. Try to build projects that solve real-world problems.</li></ol><p><strong>Tips for success</strong>:</p><p>Here are some tips to help you succeed on the exam:</p><ul><li><strong>Start early</strong> : Give yourself plenty of time to prepare for the exam. I recommend starting your preparation at least three months in advance.</li><li><strong>Be consistent</strong> : Set aside some time each day to study for the exam. Even if it’s just for an hour, consistent study will help you retain information better.</li><li><strong>Focus on the hands-on aspects of the exam :</strong> The exam will test your ability to apply your knowledge to real-world problems. Make sure you have plenty of hands-on experience with GCP services.</li><li><strong>Don’t memorize the exam guide</strong> : Instead, focus on understanding the concepts and how they apply to real-world problems.</li></ul><p><strong>Conclusion</strong> : <br>Earning the Google Cloud Machine Learning Certification in 2023 was a challenging yet rewarding journey. It not only enhanced my knowledge and skills in machine learning on the Google Cloud Platform but also opened up exciting career opportunities. Remember that preparation, dedication, and continuous learning are the keys to success in the field of machine learning and cloud computing. So if you’re considering this certification, don’t hesitate to embark on this enriching journey — the possibilities are limitless 🙂</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=15f3bdef61c1" width="1" height="1" alt=""><hr><p><a href="https://medium.com/google-developer-experts/ml-story-how-i-cleared-the-google-cloud-machine-learning-certification-in-2023-15f3bdef61c1">[ML Story] How I cleared the Google Cloud Machine Learning Certification in 2023</a> was originally published in <a href="https://medium.com/google-developer-experts">Google Developer Experts</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Video : Derivative of the Sigmoid Activation function]]></title>
            <link>https://medium.com/@bhattbhavesh91/video-derivative-of-the-sigmoid-activation-function-3c3c16318bb0?source=rss-a750aa010abc------2</link>
            <guid isPermaLink="false">https://medium.com/p/3c3c16318bb0</guid>
            <category><![CDATA[data-science]]></category>
            <category><![CDATA[neural-networks]]></category>
            <category><![CDATA[machine-learning]]></category>
            <category><![CDATA[deep-learning]]></category>
            <category><![CDATA[sigmoid]]></category>
            <dc:creator><![CDATA[Bhavesh Bhatt]]></dc:creator>
            <pubDate>Sun, 03 Nov 2019 06:21:01 GMT</pubDate>
            <atom:updated>2019-11-03T06:40:37.832Z</atom:updated>
            <content:encoded><![CDATA[<h3>Video : Derivative of the Sigmoid Activation function</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*ZQLzMrRXNRtyjcodrb-Ibw.png" /></figure><p>The sigmoid activation function is a widely used function in Deep Learning which essentially squashes values between zero and one &amp; is represented as follows.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*aJC7BlDkh2z-3YsQ" /></figure><p>Its derivative has advantageous properties, which is why it is widely used in neural networks.</p><p>The derivative of the sigmoid function looks something like this :</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*09zOmABrykQe-HpK" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*s-oj85y4gHExvkx0" /></figure><p>In this video, I’ll walk through each step of the derivation and discuss why people generally use the above mentioned form of the derivative instead of any other form.</p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fwww.youtube.com%2Fembed%2FG6djH3I0rG0%3Ffeature%3Doembed&amp;url=https%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3DG6djH3I0rG0&amp;image=https%3A%2F%2Fi.ytimg.com%2Fvi%2FG6djH3I0rG0%2Fhqdefault.jpg&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;schema=youtube" width="854" height="480" frameborder="0" scrolling="no"><a href="https://medium.com/media/2ca9c3d213ae73cef76ff0ff9b50bc33/href">https://medium.com/media/2ca9c3d213ae73cef76ff0ff9b50bc33/href</a></iframe><p>I hope you found this video informative 😊</p><p>Want to know more about me?<br>Please check out my <a href="https://bhattbhavesh91.github.io/">Website</a>.<br>If you’d like to learn machine learning, check out my <a href="https://youtube.com/bhaveshbhatt8791/">YouTube channel</a>.<br>If you’d like to get updates, follow me on <a href="https://twitter.com/_bhaveshbhatt">Twitter</a>, <a href="https://github.com/bhattbhavesh91">GitHub </a>&amp; <a href="https://medium.com/@bhattbhavesh91">Medium</a>.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=3c3c16318bb0" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[How I Got Selected as a Google Developer Expert in Machine Learning]]></title>
            <link>https://medium.com/@bhattbhavesh91/how-i-got-selected-as-a-google-developer-expert-in-machine-learning-1231381da924?source=rss-a750aa010abc------2</link>
            <guid isPermaLink="false">https://medium.com/p/1231381da924</guid>
            <category><![CDATA[google]]></category>
            <category><![CDATA[technology]]></category>
            <category><![CDATA[machine-learning]]></category>
            <category><![CDATA[data-science]]></category>
            <category><![CDATA[google-developer-expert]]></category>
            <dc:creator><![CDATA[Bhavesh Bhatt]]></dc:creator>
            <pubDate>Mon, 14 Oct 2019 12:17:16 GMT</pubDate>
            <atom:updated>2019-10-14T13:56:57.152Z</atom:updated>
            <content:encoded><![CDATA[<p>I am extremely stoked to share that as of today i.e 14th October 2019, I have been recognized by Google as a Google Developer Expert in Machine Learning.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/600/1*ylFsuLAkVXuZZ1MSfWiOiQ.png" /><figcaption>Credits: Google Developer Experts</figcaption></figure><p>I thought of writing a blog post to answer some of the questions you might have about the GDE program.</p><h3><strong>Who is a Google Developers Expert?</strong></h3><p>Google Developers Experts (GDEs) are experienced, recognized developers of Google technologies as well as outstanding professionals in product strategy, UX/UI, marketing, growth hacking and monetization. They distinguish themselves through frequently speaking at conferences, share their passion and experience by publishing videos and tutorials, writing code samples, mentoring developers and startups and much more. Thanks to their support, developers, high-potential startups and technical communities around the world build and launch highly innovative apps.</p><h3><strong>Who am I?</strong></h3><p>Before I explain the process, I’ll give you my brief background. After completing my post graduation from BITS Pilani K K Birla Goa Campus in Embedded Systems &amp; having worked for awesome companies like MathWorks &amp; Cisco, I decided to pivot my career in machine learning/data science.</p><p>In my tenure so far as a Data Scientist, I have tackled some open ended problems in the field of computer vision, time series &amp; natural language processing.</p><p>In order to give back to the community from which I learnt so much I started creating videos on YouTube and as of today, I have close to 8100 subscribers and 100 videos on my channel.</p><h3><strong>How I Got Selected as a Google Developer Expert in Machine Learning?</strong></h3><p><strong>Step 1 → Nomination <br></strong>I was nominated by a partner authorized by the Google developers team. As per the details given on the official experts portal, you can either be nominated by Google employee or any partners authorized by the Google developers team.</p><p><strong>Step 2 → Eligibility Check<br></strong>My application form was validated on factors such as speaking engagements, community engagement, mentorship, content creation, forum participation, projects made using google technologies &amp; open software contribution.</p><p><strong>Step 3 → Community Interview<br></strong>After clearing the initial eligibility check. Then, an existing GDE was assigned to interview me to assess some of my technical skills and also my impact on the developer community. The discussion was mostly centered around my YouTube videos with some questions in machine learning.</p><p><strong>Step 4 → Product Interview<br></strong>After clearing the community interview, I was interviewed by a Google specialist in machine learning. This round was technical with questions on explaining deep learning concepts in layman terms, along with some questions on machine learning basics as well as some questions on natural language processing, document similarity.</p><h3><strong>What changes after becoming a GDE in Machine Learning?</strong></h3><p>My main idea of applying for the GDE program was to teach at scale and having cleared the rounds, I’ll use this opportunity to teach machine learning at Google conferences. I’ll get a chance to showcase my machine learning projects. Additionally, I will also get a chance to experiment with the latest version of TensorFlow. The GDE role is essentially an ambassador role for Google’s Developer initiative. It’s a great opportunity to do my day job as well as be involved with the wider machine learning community.</p><h3>A few tips</h3><p>Don’t apply for the tag, apply for the evangelism. GDE’s didn’t become GDE’s with the aim of becoming GDE’s. Being a GDE is really about giving back to the community. Find something that you enjoy doing (writing blogs, publishing books, creating content, public speaking) and how you can help others.</p><p>I hope you found this blog informative 😊</p><p>Want to know more about me? <br>Please check out my <a href="https://bhattbhavesh91.github.io/">Website</a>. <br>If you’d like to learn machine learning, check out my <a href="https://youtube.com/bhaveshbhatt8791/">YouTube channel</a>.<br>If you’d like to get updates, follow me on <a href="https://twitter.com/_bhaveshbhatt">Twitter</a>, <a href="https://github.com/bhattbhavesh91">GitHub </a>&amp; <a href="https://medium.com/@bhattbhavesh91">Medium</a>.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=1231381da924" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The mystery of True Positive, True Negative, False Positive and False Negative]]></title>
            <link>https://medium.com/@bhattbhavesh91/the-mystery-of-true-positive-true-negative-false-positive-and-false-negative-fd73c78c905a?source=rss-a750aa010abc------2</link>
            <guid isPermaLink="false">https://medium.com/p/fd73c78c905a</guid>
            <category><![CDATA[cricket]]></category>
            <category><![CDATA[data-science]]></category>
            <category><![CDATA[statistics]]></category>
            <category><![CDATA[machine-learning]]></category>
            <dc:creator><![CDATA[Bhavesh Bhatt]]></dc:creator>
            <pubDate>Sun, 10 Sep 2017 11:09:43 GMT</pubDate>
            <atom:updated>2019-10-25T03:12:23.126Z</atom:updated>
            <content:encoded><![CDATA[<p>I’m sure most of you are always confused regarding when an event is True Positive, True Negative, False Positive and False Negative.</p><p>I am using cricket the sport to explain this simple concept.</p><p><strong>True Positive</strong></p><ul><li>Umpire gives a Batsman <strong>NOT OUT </strong>when he is <strong>NOT OUT.</strong></li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/574/1*3n8XHTFenXglyAMLC4uZ-Q.jpeg" /></figure><p><strong>True Negative</strong></p><ul><li>Umpire gives a Batsman <strong>OUT </strong>when he is <strong>OUT.</strong></li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*FZdK99Psf7bmoOUBbZAEoQ.jpeg" /></figure><p><strong>False Positive (Type I error : Hurts a bit)</strong></p><ul><li>Umpire gives a Batsman <strong>NOT OUT </strong>when he is <strong>OUT.</strong></li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/904/1*6xbtCNNfb1bM3lijeQ7vmw.jpeg" /></figure><p><strong>False Negative (Type II error : Hurts the most)</strong></p><ul><li>Umpire gives a Batsman <strong>OUT </strong>when he is <strong>NOT OUT.</strong></li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/900/1*9zpqZ8O0e-mEvEW-ecrT3A.jpeg" /></figure><p>I hope it helps in demystifying the mystery!</p><p>Want to know more about me?<br>Please check out my <a href="https://bhattbhavesh91.github.io/">Website</a>.<br>If you’d like to learn machine learning, check out my <a href="https://youtube.com/bhaveshbhatt8791/">YouTube channel</a>.<br>If you’d like to get updates, follow me on <a href="https://twitter.com/_bhaveshbhatt">Twitter</a>, <a href="https://github.com/bhattbhavesh91">GitHub </a>&amp; <a href="https://medium.com/@bhattbhavesh91">Medium</a>.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/351/1*h9LgwWW3zSlirux2N4gv4w.jpeg" /></figure><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=fd73c78c905a" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Recommendation system implementation for Dummies]]></title>
            <link>https://medium.com/@bhattbhavesh91/recommendation-system-implementation-for-dummies-9051508b4398?source=rss-a750aa010abc------2</link>
            <guid isPermaLink="false">https://medium.com/p/9051508b4398</guid>
            <category><![CDATA[banking]]></category>
            <category><![CDATA[data-science]]></category>
            <category><![CDATA[recommender-systems]]></category>
            <category><![CDATA[recommendation-system]]></category>
            <category><![CDATA[machine-learning]]></category>
            <dc:creator><![CDATA[Bhavesh Bhatt]]></dc:creator>
            <pubDate>Sun, 27 Aug 2017 09:30:50 GMT</pubDate>
            <atom:updated>2017-08-27T09:30:50.502Z</atom:updated>
            <content:encoded><![CDATA[<ul><li>Imagine that you have recently gone bankrupt and 1 fine day you get a call from a bank agent requesting you to subscribe for a term deposit.</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/1002/1*I0_3Y5-56rtAxAi8XTpXtg.jpeg" /></figure><p>This is the reaction most of us will have.</p><ul><li>Banks too invest a lot, they invest in the agents that call you up and also the telephone bills are huge.</li></ul><p>What if I tell you, both the end user and the bank can be profited by using a simple recommendation system.</p><p>I’ll describe the approach below.</p><p>For the demo purpose, I have used the bank dataset from the below link -</p><p><a href="https://archive.ics.uci.edu/ml/datasets/bank+marketing">https://archive.ics.uci.edu/ml/datasets/bank+marketing</a></p><p>The bank already has data of 42000+ customers with the following details<br>- age, job, marital status, education, loan defaulter, balance, housing loan, personal loan.<br>- last contacted, day, month, duration, previous outcome<br>- Finally variable y whether you did subscribe for the term plan or not.</p><p>After doing a bit of preprocessing to the data</p><ul><li>Converting Yes and No to 0 and 1.</li><li>Converting Categorical variables to Numeric variables using One Hot encoding.</li><li>Dropping columns such as contact, day and month which typically don’t contribute much to the prediction of the target variable.</li></ul><p>This is something that I get after doing a bit of preprocessing.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*tJ7cug4yx9lRe5pkTMAkSw.png" /></figure><p>I split my data into train and test and predict the outcome based on logistic regression model.</p><p>I get an accuracy score of around 90%</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/998/1*2yNt9iCqsBTpOXXn2J8v2w.png" /></figure><p>My goal isn’t however solved. What is the use of a tool if I cannot use it in my data to data life.</p><p>Suppose a new customer comes in with the following details</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/342/1*JxjCNJ2A7kugvnDkwvE5Rw.png" /></figure><p>I just call my recommendation engine and check the output and what do I observe?</p><p>Any guesses? Will I be telling my customer agent to call this guy up for a term deposit subscription?</p><p>The recommendation engine gives me the following output</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/165/1*-t-_4nLQNrY5VgNG67WPAQ.png" /></figure><p>It gives me an output 0 with a probability of 0.96!</p><p>So, I can clearly say that this person won’t be interested for a term deposit subscription based on the data he has provided.</p><ul><li>Its a 2 way benefit scheme, the banks can target users which they now feel can subscribe to a term deposit based on the data and this will also reduce the workforce and the budget required for the sales operation of such a term deposit subscription.</li><li>The users who are genuinely in need of term deposit subscription will be contacted and the other can work hard to enroll themselves in the coming future for a term deposit subscription.</li><li>For complete code visit : <a href="https://github.com/bhattbhavesh91/Term_Deposit_Recommendor_System/blob/master/Term_Deposite_Recommendor_System.ipynb">https://github.com/bhattbhavesh91/Term_Deposit_Recommendor_System/blob/master/Term_Deposite_Recommendor_System.ipynb</a></li></ul><p>I know it was this simple</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/206/1*lv4rb-kib0Beu0qtWPkm2Q.jpeg" /></figure><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=9051508b4398" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The mystery of n_components in PCA]]></title>
            <link>https://medium.com/@bhattbhavesh91/the-mystery-of-n-components-in-pca-a0431ee995be?source=rss-a750aa010abc------2</link>
            <guid isPermaLink="false">https://medium.com/p/a0431ee995be</guid>
            <category><![CDATA[data-science]]></category>
            <category><![CDATA[machine-learning]]></category>
            <category><![CDATA[python]]></category>
            <category><![CDATA[algorithms]]></category>
            <dc:creator><![CDATA[Bhavesh Bhatt]]></dc:creator>
            <pubDate>Fri, 25 Aug 2017 18:08:41 GMT</pubDate>
            <atom:updated>2017-08-27T02:35:09.741Z</atom:updated>
            <content:encoded><![CDATA[<ul><li>We often stop valuing things when they are given to us readily.</li><li>Something similar happens to us while solving a Machine Learning problem. scikit-learn pampers you so much that you stop thinking how an algorithm really functions.</li><li>One fine day, randomly someone asked me if you have a feature matrix of consisting of 20 features and if you plan to reduce the dimension, what is the n_components you will chose or how many optimum features will you use to represent the 20 dimensional data without having data loss?</li></ul><p>So the best that you can do when you are clueless is smile and try to escape the situation which is what I did :)</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/450/1*IEWIZD3T05mrxVm2GAVzhA.jpeg" /></figure><p>So, the question still remains, how do you find the right number of features to represent a data set consisting of 20 features.</p><p>So, I decided to get my hands dirty!</p><p>Thought of starting off with the famous Iris Dataset!</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*4wI2i7fqMRYPT2XFQAVFjw.png" /></figure><p>Thank god those flowers were made and someone took so much amount of effort to record the length and width of the petals and sepals. Without the Iris data set Machine learning would have been so difficult.</p><ul><li>It has 4 features and 150 entries.</li></ul><p>PCA is as simple as the number of letters in PCA</p><ol><li>Compute the co-variance matrix</li><li>Compute the eigen values and vectors.</li><li>Project the data along the top eigen vectors based on the n_components.</li></ol><p>Thats it, the job is done! But wait,</p><p>How do we reach to n_components??</p><p>So let me solve the mystery</p><ul><li>Compute the eigen values and vectors</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/749/1*LiJDwuLm4u_nBJceoi226g.png" /></figure><ul><li>Arrange the eigen values in the descending order.</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/991/1*5QiKZzfJlWI0hvta8QXACA.png" /></figure><ul><li>Since eigen values capture the variance by each component in the direction of the eigen vector. See the percentage of variance contributed by each feature which in turn contributes to the predicting power.</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*M3W9MSBE2AG0Z1sqY8_X0g.png" /></figure><p>Clearly the variance captured by the 1st 2 features when arranged in descending order contribute to almost 96% of the total variance by the features. Thus we can eliminate the remaining 2 features as they don’t contribute much to the overall variance.</p><p>Thus, the mystery is solved.</p><p>n_components should be equal to the features which contribute a large number to the overall variance! The number depends on the business logic.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/550/1*PlCde9j8gcPH1QIPbdYLLQ.jpeg" /></figure><p>For the complete code, please visit -</p><p><a href="https://github.com/bhattbhavesh91/pca_from_scratch_iris_dataset/blob/master/Principal_Component_Analysis_Iris_DataSet.ipynb">https://github.com/bhattbhavesh91/pca_from_scratch_iris_dataset/blob/master/Principal_Component_Analysis_Iris_DataSet.ipynb</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=a0431ee995be" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[SQL Joins Simplified]]></title>
            <link>https://medium.com/@bhattbhavesh91/sql-joins-simplified-f2974eddea29?source=rss-a750aa010abc------2</link>
            <guid isPermaLink="false">https://medium.com/p/f2974eddea29</guid>
            <category><![CDATA[machine-learning]]></category>
            <category><![CDATA[sql]]></category>
            <category><![CDATA[data-science]]></category>
            <category><![CDATA[programming]]></category>
            <dc:creator><![CDATA[Bhavesh Bhatt]]></dc:creator>
            <pubDate>Mon, 21 Aug 2017 16:52:30 GMT</pubDate>
            <atom:updated>2017-08-21T16:59:45.752Z</atom:updated>
            <content:encoded><![CDATA[<p>A SQL join is a Structured Query Language (SQL) instruction to combine data from two sets of data (e.g. two tables).</p><p>Lets first create 2 tables to join</p><blockquote>create table t1(cid int(20), cname char(20));</blockquote><blockquote>insert into t1 values<br>(1,’India’), <br>(2, ‘USA’), <br>(3, ‘Australia’);</blockquote><p>Table 1 = <br>+ — — — + — — — — — -+<br>| cid | cname |<br>+ — — — + — — — — — -+<br>| 1 | India |<br>| 2 | USA |<br>| 3 | Australia |<br>+ — — — + — — — — — -+</p><p>Lets create a 2nd table</p><blockquote>create table t2(sid int(20), cid int(20), sname char(20));</blockquote><blockquote>insert into t2 values<br>(1,1,’Maharashtra’),<br> (2,1,’Punjab’),<br> (3,2,’Massachusetts’),<br> (4,NULL,’Kathmandu’);</blockquote><p>This is how our 2nd table looks -</p><p>Table 2 = <br>+ — — — + — — — + — — — — — — — -+<br>| sid | cid | sname |<br>+ — — — + — — — + — — — — — — — -+<br>| 1 | 1 | Maharashtra |<br>| 2 | 1 | Punjab |<br>| 3 | 2 | Massachusetts |<br>| 4 | NULL | Kathmandu |<br>+ — — — + — — — + — — — — — — — -+</p><h3><strong>Lets start with the Inner Join</strong></h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/300/1*6kTqnnbSx45O0Zd2wub86g.png" /></figure><p>Inner join</p><blockquote>mysql&gt; select * from t1 inner join t2 on t1.cid = t2.cid;</blockquote><ul><li>— — — + — — — -+ — — — + — — — + — — — — — — — -+<br>| cid | cname | sid | cid | sname |<br>+ — — — + — — — -+ — — — + — — — + — — — — — — — -+<br>| 1 | India | 1 | 1 | Maharashtra |<br>| 1 | India | 2 | 1 | Punjab |<br>| 2 | USA | 3 | 2 | Massachusetts |<br>+ — — — + — — — -+ — — — + — — — + — — — — — — — -+</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/328/1*2qbhjG60O6u2H8sSzzr-ow.png" /></figure><p><strong>Left join</strong></p><blockquote>mysql&gt; select * from t1 left join t2 on t1.cid = t2.cid;</blockquote><ul><li>— — — + — — — — — -+ — — — + — — — + — — — — — — — -+<br>| cid | cname | sid | cid | sname |<br>+ — — — + — — — — — -+ — — — + — — — + — — — — — — — -+<br>| 1 | India | 1 | 1 | Maharashtra |<br>| 1 | India | 2 | 1 | Punjab |<br>| 2 | USA | 3 | 2 | Massachusetts |<br>| 3 | Australia | NULL | NULL | NULL |<br>+ — — — + — — — — — -+ — — — + — — — + — — — — — — — -+</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/303/1*TBrB6Rxt_Vr2nFFndxgBPg.png" /></figure><p>Right join</p><blockquote>mysql&gt; select * from t1 right join t2 on t1.cid = t2.cid;</blockquote><ul><li>— — — + — — — -+ — — — + — — — + — — — — — — — -+<br>| cid | cname | sid | cid | sname |<br>+ — — — + — — — -+ — — — + — — — + — — — — — — — -+<br>| 1 | India | 1 | 1 | Maharashtra |<br>| 1 | India | 2 | 1 | Punjab |<br>| 2 | USA | 3 | 2 | Massachusetts |<br>| NULL | NULL | 4 | NULL | Kathmandu |<br>+ — — — + — — — -+ — — — + — — — + — — — — — — — -+</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/322/1*2sgYll-IFKR0OxiSzR75hA.png" /></figure><blockquote>mysql&gt; select * from t1 left join t2 on t1.cid = t2.cid union select * from t1 right join t2 on t1.cid = t2.cid;</blockquote><p>+ — — — + — — — — — -+ — — — + — — — + — — — — — — — -+<br>| cid | cname | sid | cid | sname |<br>+ — — — + — — — — — -+ — — — + — — — + — — — — — — — -+<br>| 1 | India | 1 | 1 | Maharashtra |<br>| 1 | India | 2 | 1 | Punjab |<br>| 2 | USA | 3 | 2 | Massachusetts |<br>| 3 | Australia | NULL | NULL | NULL |<br>| NULL | NULL | 4 | NULL | Kathmandu |<br>+ — — — + — — — — — -+ — — — + — — — + — — — — — — — -+</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=f2974eddea29" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Intuitive explanation of the Central Limit Theorem]]></title>
            <link>https://medium.com/@bhattbhavesh91/intuitive-explanation-of-the-central-limit-theorem-e11027b5c2ce?source=rss-a750aa010abc------2</link>
            <guid isPermaLink="false">https://medium.com/p/e11027b5c2ce</guid>
            <category><![CDATA[statistics]]></category>
            <dc:creator><![CDATA[Bhavesh Bhatt]]></dc:creator>
            <pubDate>Sat, 01 Jul 2017 08:54:44 GMT</pubDate>
            <atom:updated>2017-07-01T08:54:44.123Z</atom:updated>
            <content:encoded><![CDATA[<h3>What is the ‘Central Limit Theorem — CLT’</h3><p>The central limit theorem (CLT) is a statistical theory that states that given a sufficiently large sample size from a population with a finite level of variance, the mean of all samples from the same population will be approximately equal to the mean of the population. Furthermore, all of the samples will follow an approximate normal distribution pattern, with all variances being approximately equal to the variance of the population divided by each sample’s size.</p><p>Imagine rolling a 1-6 die lots of times. We expect to get (over a long time) an equal proportion of each roll.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/600/1*kf-uHpw5Z8oyc8ymXjcnUg.png" /></figure><p>Next, do the same thing with 2 dice. You only have a 1 in 36 (around 0.03) chance of a 2 or a 12 because there’s only 1 way to make a 2 — you need a 1 on both dice. But a seven is much more common — you have a 1 in 6 (around 0.16) chance of a 7.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/600/1*R95-uEr_XEHg96pBMIh3og.png" /></figure><p>Now let’s do it again, this time with three dice. Chance of a 3 (or an 18) is 1 in 216 (6*6*6). Chance of a 10 or 11 is much higher.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/600/1*FEMW7MHKlF5BITnOebueGA.png" /></figure><p>But the shape of the distribution has gone from flat (uniform) to something that looks pretty close to a normal distribution, and we’ve only got three dice. Each time you add a die, the distribution gets closer to normal. And that’s because of central limit theorem.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=e11027b5c2ce" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Co-variance and Correlation]]></title>
            <link>https://medium.com/@bhattbhavesh91/co-variance-and-correlation-b4597237a4e7?source=rss-a750aa010abc------2</link>
            <guid isPermaLink="false">https://medium.com/p/b4597237a4e7</guid>
            <category><![CDATA[correlation]]></category>
            <category><![CDATA[data-science]]></category>
            <dc:creator><![CDATA[Bhavesh Bhatt]]></dc:creator>
            <pubDate>Fri, 30 Jun 2017 10:16:01 GMT</pubDate>
            <atom:updated>2017-06-30T10:24:18.097Z</atom:updated>
            <content:encoded><![CDATA[<p><strong>Co-variance</strong> between two variables is positive when they tend to move in the same direction and negative if they tend to move in opposite directions.</p><p>To see how we arrive at the definition, first think about two variables each with mean zero. That is, each variable could be positive or negative but is on average zero. To summarize whether the two variables move together, we can look at the product:</p><ul><li>When both variables have the same sign, it’s positive;</li><li>When they have opposite signs, it’s negative;</li><li>When they are both large and have the same sign, you get a big positive number;</li><li>When they are both large and have opposite signs, you get a big negative number;</li></ul><p>The average value of this product is the co-variance: it’s positive when the two variables tend to have the same sign and negative when they tend to have opposite signs (weighting by magnitude).</p><p>What if the two variables do not have mean zero? Well, the means aren’t relevant to whether the variables move together or not. The mean only describes the overall level of each. So we don’t care about the means for this purpose, and to make things easier, we may as well set them both to zero (by subtracting off the real mean of each). Then we can look at the product as above.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/526/1*5r1B3ni7A4JX8jziFRSwaw.png" /></figure><p><strong>Correlation</strong> is a scaled version of co-variance that takes on values in [−1,1] with a correlation of ±1 indicating perfect linear association and 0 indicating no linear relationship. This scaling makes correlation invariant to changes in scale of the original variables.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/723/1*Ma8WwjwfWf_EDGe7u3631g.png" /></figure><p>I decided to get my hands dirty by trying to plot a relation between temperature and ice cream sales using pandas.</p><pre>import numpy as np<br>import pandas as pd</pre><pre>temperature = [66,72,77,84,83,71,65,70]<br>ice_cream_sales = [8,11,15,20,21,11,8,10]<br>my_data_frame = pd.DataFrame(list(zip(temperature,ice_cream_sales)),columns=[&#39;Temperature&#39;, &#39;Icecream Sales&#39;])</pre><figure><img alt="" src="https://cdn-images-1.medium.com/max/284/1*KXvV4Bp6WHSWnYUPrWWiXA.png" /></figure><p>Co-variance is calculated by -</p><pre>my_data_frame.cov()</pre><figure><img alt="" src="https://cdn-images-1.medium.com/max/382/1*6FsxWIfBN4UEfiCYOutTHg.png" /></figure><p>Correlation is calculated by -</p><pre>my_data_frame.corr()</pre><figure><img alt="" src="https://cdn-images-1.medium.com/max/386/1*S4tVSwJiXU45nG_Sdy_dFw.png" /></figure><p>As we can clearly see, 0.988046 clearly denotes that there is a strong relation between Ice cream sales and Temperature.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=b4597237a4e7" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>