<?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 Raghu Nandhan Meegada on Medium]]></title>
        <description><![CDATA[Stories by Raghu Nandhan Meegada on Medium]]></description>
        <link>https://medium.com/@meegadaraghuyadhav?source=rss-eb90b683e7be------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*WWcZ4-OfzH-2FyO5sfeXiQ.png</url>
            <title>Stories by Raghu Nandhan Meegada on Medium</title>
            <link>https://medium.com/@meegadaraghuyadhav?source=rss-eb90b683e7be------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Sun, 22 Nov 2026 20:38:21 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@meegadaraghuyadhav/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[Beyond the Syntax: Unveiling the Hidden Logic of SQL Query Execution]]></title>
            <link>https://medium.com/@meegadaraghuyadhav/beyond-the-syntax-unveiling-the-hidden-logic-of-sql-query-execution-a7e7b1bae751?source=rss-eb90b683e7be------2</link>
            <guid isPermaLink="false">https://medium.com/p/a7e7b1bae751</guid>
            <dc:creator><![CDATA[Raghu Nandhan Meegada]]></dc:creator>
            <pubDate>Sat, 21 Jun 2025 13:34:37 GMT</pubDate>
            <atom:updated>2025-06-21T13:34:37.830Z</atom:updated>
            <content:encoded><![CDATA[<h3>Discover what really happens beneath your SQL code — step by step, clause by clause.</h3><p>When you write a SQL query, you naturally start with the SELECT clause and then proceed with FROM, WHERE, etc. But what if I told you that the database doesn’t execute the query in that order at all?</p><p>Understanding the <strong>logical execution order</strong> of SQL is <strong>critical for performance optimization, debugging, and writing accurate queries</strong> — yet it’s often overlooked by beginners and even intermediates.</p><p>This article dives deep into the actual order in which SQL queries are executed, provides uncommon insights, and offers practical examples using real syntax and best practices. Let’s unveil the truth.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1018/1*dpIVIARM6qT6zfxnUw5FRg.png" /><figcaption>SQL Order of Execution</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/507/1*BDweTFSJbxBDESmICwG1Hg.png" /><figcaption>creating database &amp; describing employees table datatypes</figcaption></figure><p><strong>Tip</strong>: The engine doesn’t “see” SELECT until all filters and grouping are done!</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*HcSCj_XeutE9oUy6PPhdWg.png" /><figcaption>Table of Clause &amp; its Purposes</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/485/1*1JFNRjg9yUp8LL-grsahLA.png" /></figure><p><strong>Execution Flow:</strong></p><p>1.<strong>FROM employees</strong> — Get all rows from employees.</p><p>2.<strong>WHERE active = 1</strong> — Retain only active employees.</p><p>3.<strong>GROUP BY department</strong> — Aggregate by department.</p><p>4.<strong>HAVING AVG(salary) &gt; 60000</strong> — Filter only departments with high average salaries.</p><p>5.<strong>SELECT department, AVG(salary)</strong> — Choose relevant columns.</p><p>6.<strong>ORDER BY avg_sal DESC</strong> — Sort by average salary.</p><p>7.<strong>LIMIT 5</strong> — Show only top 5.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/725/1*H6EBJA0tiHXMN5Y3Wjzrzg.png" /></figure><h3><strong>Hidden Insights &amp; Gotchas</strong></h3><p><strong>1. HAVING is <em>not</em> a replacement for WHERE</strong></p><p>•WHERE filters <strong>rows</strong> before aggregation.</p><p>•HAVING filters <strong>groups</strong> after aggregation.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/782/1*fUgcZryS6cTTS_DySS_acw.png" /></figure><p>💡 <strong>Unknown Insight</strong>: When you use WHERE on the joined table (o), the NULLs introduced by LEFT JOIN are filtered out, effectively acting like an INNER JOIN.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*hBLjZc0D9QIo9IsG65CRGA.png" /></figure><h3><strong>4. The Myth of SELECT *</strong></h3><p>Using SELECT * seems innocent but hides performance costs.</p><p>•Brings unnecessary columns → more I/O</p><p>•Breaks when schema changes</p><p>•Causes ambiguity in joins</p><p>👎 Avoid SELECT * in production queries.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/525/1*na4zoRVep2X4x-OiouRMBw.png" /></figure><p>✍️ <em>Article authored by Raghu Nandhan !!</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=a7e7b1bae751" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The Silent Architect of Intelligence: Data Preprocessing in Machine Learning]]></title>
            <link>https://medium.com/@meegadaraghuyadhav/the-silent-architect-of-intelligence-data-preprocessing-in-machine-learning-568334dc1ca4?source=rss-eb90b683e7be------2</link>
            <guid isPermaLink="false">https://medium.com/p/568334dc1ca4</guid>
            <dc:creator><![CDATA[Raghu Nandhan Meegada]]></dc:creator>
            <pubDate>Tue, 27 May 2025 17:25:17 GMT</pubDate>
            <atom:updated>2025-05-27T17:25:17.340Z</atom:updated>
            <content:encoded><![CDATA[<p><strong><em>By Raghu Nandhan Meegada ,</em></strong><em>B.Tech in Artificial Intelligence,</em><br><em>Published on Medium Digest — AI Research Series</em></p><p><strong><em>“Better data beats clever algorithms.” — Pedro Domingos</em></strong></p><p><em>“You cannot build a palace on sand. Likewise, you cannot build intelligence on raw, chaotic data.”</em></p><p><strong>Introduction:</strong></p><p><strong>The Unseen Engine: Why Data Preprocessing is Crucial in Machine Learning</strong></p><p>In the grand cathedral of machine learning, most attention is drawn to the visible spectacle — neural networks, random forests, transformers. Yet behind these elegant algorithms, there stands a silent architect, often unnoticed but always vital: data preprocessing.</p><p>This phase does not merely “prepare” data. It is the act of understanding it. In the age of Artificial Intelligence, <strong>data is the new oil</strong>, but raw oil is not valuable until refined. Likewise, raw data is often incomplete, noisy, inconsistent, and unwieldy. Without proper treatment, even the most sophisticated machine learning models can crumble under the weight of poor data.</p><p>This brings us to a critical but often underappreciated phase in the ML pipeline: Data Preprocessing</p><figure><img alt="https://realpython.com/learning-paths/classic-data-structures-and-algorithms-with-python/" src="https://cdn-images-1.medium.com/max/1024/1*SThlGOU6wKlGRRK69Rmfhw.jpeg" /></figure><p><strong>The Nature of Raw Data: Uncivilized, Unforgiving</strong></p><p>Data, in its natural form, is not generous. It is often:</p><ul><li>Incomplete (missing entries)</li><li>Incoherent (inconsistent formats)</li><li>Noisy (random variations, typos)</li><li>Biased (unbalanced representation)</li><li>Redundant (duplicate or irrelevant features)</li></ul><p>Take a hospital database. Some patients are missing blood pressure readings. Others have “M” or “Male” inconsistently written. One variable might record weight in kilograms, another in pounds. Feeding such data into an algorithm is like giving a chef rotten ingredients and expecting a gourmet dish.</p><p>Thus, before algorithms can learn, data must be taught to speak — in a language they can understand.</p><figure><img alt="https://medium.com/microsoft-power-bi/tagged/data" src="https://cdn-images-1.medium.com/max/1024/1*Jn-l0Bajln_P9J9BorsnHQ.png" /></figure><p><strong>Data Preprocessing: The Ritual of Refinement</strong></p><p>Data preprocessing is not a checklist; it is a dialogue between the analyst and the dataset. Each step answers a question:</p><ol><li><strong>What do you know?</strong><br> <em>Cleaning</em>: Handle missing values, correct anomalies, remove duplicates.</li><li><strong>What are you?</strong><br> <em>Transformation</em>: Encode categorical data, scale numerical features, normalize distributions.</li><li><strong>What do you mean?</strong><br> <em>Feature Engineering</em>: Extract meaning, build interactions, derive new attributes.</li><li><strong>What matters most?</strong><br> <em>Feature Selection &amp; Reduction</em>: Drop the irrelevant, distill the essence.</li><li><strong>What should I trust?</strong><br> <em>Splitting and validation</em>: Protect against bias, data leakage, and overfitting.</li></ol><figure><img alt="" src="https://cdn-images-1.medium.com/max/833/1*yCsEZhasEywu8oRj5lglww.png" /></figure><p>These aren’t mechanical steps — they are acts of interpretation. Each transformation embeds <strong>assumptions, domain knowledge, and logic</strong> into the dataset, turning it from numbers into meaning.</p><p><strong>Why It Matters: The Butterfly Effect of Data Quality</strong></p><p>Imagine two students learning from different textbooks. One has a clear, concise book with well-structured chapters. The other has pages missing, incorrect diagrams, and jumbled sentences. Which student will learn better?</p><p>Machine learning algorithms are those students. <strong>Their “intelligence” is shaped not by code, but by the clarity of the data.</strong></p><p><strong>Real-World Effects</strong></p><ul><li>A cancer diagnosis model may misclassify a malignant tumor as benign if input features are poorly scaled.</li><li>A fraud detection system might flag the wrong transaction because country codes were inconsistently formatted.</li><li>A language model might fail to understand sentiment because stop words weren’t removed or tokenization</li></ul><p>In each case, the failure isn’t in the algorithm — it’s in the unseen errors baked into the input.</p><p>Matching the original images impact of model building</p><figure><img alt="https://www.mdpi.com/1999-4893/17/3/128" src="https://cdn-images-1.medium.com/max/850/1*l0-iSmC5yvVxnu5bg3l5Iw.png" /></figure><p><strong>The Hidden Wisdom: Philosophical Insights</strong></p><p>Preprocessing teaches us something profound: <strong>learning is only as good as perception</strong>.</p><p>Just like human intelligence depends on what we see, hear, or experience — machine intelligence depends on the data we allow it to “feel.” Clean data is not just technically better; it is epistemologically richer — it carries clearer truths.</p><p>Furthermore, preprocessing is where <strong>bias is either corrected or reinforced</strong>. Gender imbalances, racial underrepresentation, geographic skew — these can be diluted or magnified based on how we handle the data. In that sense, preprocessing is not just a technical step, but an ethical one.</p><p>Preprocessing is <strong>not a chore</strong> — it is a craft.</p><p><strong>Conclusion: The Foundation Beneath the Future</strong></p><p>Machine learning promises great things: self-driving cars, precision medicine, smarter cities. But none of this is possible if the data underneath is cracked.</p><p>Data preprocessing may be invisible, but it is <strong>the backbone of every intelligent system</strong>. It ensures that algorithms do not just process information — they understand it.</p><p>In the end, intelligence — whether human or artificial — doesn’t emerge from brilliance alone. It emerges from clarity. And clarity begins with the data.</p><figure><img alt="https://ijrpr.com/uploads/V4ISSUE11/IJRPR19014.pdf" src="https://cdn-images-1.medium.com/max/1024/1*XMZB8z2Jkjv7Z1k350cODg.jpeg" /></figure><p><strong>References</strong></p><ul><li>Domingos, P. (2012). A Few Useful Things to Know About Machine Learning. <em>Communications of the ACM</em>, 55(10), 78–87.</li><li>Müller, A. C., &amp; Guido, S. (2017). <em>Introduction to Machine Learning with Python: A Guide for Data Scientists</em>. O’Reilly Media.</li><li>Shalev-Shwartz, S., &amp; Ben-David, S. (2014). <em>Understanding Machine Learning: From Theory to Algorithms</em>. Cambridge University Press.</li></ul><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=568334dc1ca4" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Constructing Intelligence: A Grounded Approach to the Foundations and Frontiers of Machine Learning]]></title>
            <link>https://medium.com/@meegadaraghuyadhav/constructing-intelligence-a-grounded-approach-to-the-foundations-and-frontiers-of-machine-learning-3b76a2c35e08?source=rss-eb90b683e7be------2</link>
            <guid isPermaLink="false">https://medium.com/p/3b76a2c35e08</guid>
            <dc:creator><![CDATA[Raghu Nandhan Meegada]]></dc:creator>
            <pubDate>Tue, 27 May 2025 13:34:02 GMT</pubDate>
            <atom:updated>2025-05-27T13:34:02.182Z</atom:updated>
            <content:encoded><![CDATA[<p><strong><em>By Raghu Nandhan Meegada, </em></strong><em>B.Tech in Artificial Intelligence, </em><br> <em>Published on Medium Digest — AI Research Series</em></p><p><strong>Abstract</strong></p><p>Machine Learning (ML) has emerged as the cornerstone of modern computational intelligence. No longer confined to theory, its impact is visible across every data-driven domain — from genomics and finance to autonomous vehicles and large language models.</p><p><strong>Demystifying Machine Learning: A Fresh Lens on an Ubiquitous Technology</strong></p><p>Machine learning is no longer confined to academic labs or science fiction — it’s now the invisible engine that powers our modern digital existence. From Spotify song suggestions and Netflix movie picks to email spam filters and self-driving cars, machine learning (ML) is everywhere. Yet, beneath this widespread application lies a beautifully structured framework that is as intellectually profound as it is practically impactful.</p><p>In this article, we explore machine learning not through the usual buzzwords, but through a fresh, foundational lens that uncovers how and why it works — highlighting lesser-discussed but critical concepts in the process.</p><figure><img alt="https://www.google.com/url?sa=i&amp;url=https%3A%2F%2Fx.com%2Fabner_wang_&amp;psig=AOvVaw3CG9nAAwxGEM_-0Gr7o1lP&amp;ust=1748437664336000&amp;source=images&amp;cd=vfe&amp;opi=89978449&amp;ved=0CBMQjhxqFwoTCPDX6oHcw40DFQAAAAAdAAAAABAL" src="https://cdn-images-1.medium.com/max/700/1*QQHIc-QDFM6ManvDCYzGDw.jpeg" /></figure><p><strong>1. Introduction</strong></p><p>At its essence, <strong>Machine Learning is the art and science of learning patterns from data to make predictions or decisions without being explicitly programmed.</strong> It resides at the intersection of:</p><ul><li><strong>Statistical inference</strong></li><li><strong>Optimization theory</strong></li><li><strong>Computational complexity</strong></li><li><strong>Information theory</strong></li></ul><p>As Arthur Samuel defined it in 1959:</p><p>“Machine Learning is the field of study that gives computers the ability to learn without being explicitly programmed.”</p><p>Today, ML powers the recommendation engine of Netflix, the fraud detection system of Mastercard, and even the diagnostic systems used in oncology labs.</p><figure><img alt="https://www.google.com/search?sca_esv=4ab96c2739f14cca&amp;sxsrf=AE3TifPQAjQQcMO8KRZ3UPzcCuN2e5YmpA:1748351537814&amp;q=introduction+to+machine+learning&amp;udm=2&amp;fbs=AIIjpHxU7SXXniUZfeShr2fp4giZ1Y6MJ25_tmWITc7uy4KIeoJTKjrFjVxydQWqI2NcOhZVmrJB8DQUK5IzxA2fZbQF4YL5sNSRJGgx0e9Z9AxEx3vmVdfiFyi5YpIFVpd9n752R3At7er80XW4kxbM0Pguk8Nk_GP_cwFjCfHF2otavMg4Xwl7sZtdbm54cRXfTKHNHGXpQZ-pKcwp9maStDs-5icYaQ&amp;sa=X&amp;ved=2ahUKEwivl-qC3cONAxX0kq8BHRaIGVYQtKgLegQIDRAB&amp;biw=1600&amp;bih=747&amp;dpr=1.2" src="https://cdn-images-1.medium.com/max/921/1*y2PWMk82OjE8Kd4Zb1KRNw.png" /></figure><p><strong>2. Why Machine Learning?</strong></p><p><strong>The Shift From Rule-Based to Data-Driven Systems</strong></p><p>Historically, intelligent systems were designed using handcrafted logic:</p><pre><br>def is_spam(email):<br>    spam_keywords = [&#39;free&#39;, &#39;win&#39;, &#39;urgent&#39;]<br>    return any(word in email.lower() for word in spam_keywords</pre><p>But such systems:</p><ul><li>Do not generalize</li><li>Fail with edge cases Are brittle and domain-specific</li></ul><p>Instead, ML systems learn from <em>data</em>, <strong>discovering rules automatically</strong> and <strong>adapting to changes in distribution</strong>.</p><figure><img alt="https://www.google.com/search?q=introductionmachine+learningimages+cartoon&amp;sca_esv=4ab96c2739f14cca&amp;udm=2&amp;biw=1600&amp;bih=747&amp;sxsrf=AE3TifNz7bsz_tBWnEokeJAKUC6QgUgOkw%3A1748350632718&amp;ei=qLY1aNPNK8mPseMPvrjg8QE&amp;ved=0ahUKEwjTwZ_T2cONAxXJR2wGHT4cOB4Q4dUDCBE&amp;uact=5&amp;oq=introductionmachine+learningimages+cartoon&amp;gs_lp=EgNpbWciKmludHJvZHVjdGlvbm1hY2hpbmUgbGVhcm5pbmdpbWFnZXMgY2FydG9vbkj7KFCFAliBJ3ABeACQAQCYAXagAe4PqgEEMTYuNbgBA8gBAPgBAZgCCaAC2QbCAgcQIxgnGMkCmAMA4gMFEgExIECIBgGSBwM3LjKgB7cnsgcDNy4yuAfZBsIHB" src="https://cdn-images-1.medium.com/max/940/1*9vZjG-vR3EMgFB0YjAZHoQ.png" /></figure><p><strong>Generalization</strong></p><p>The power of ML lies in generalizing from observed data to unseen instances. This generalization capability is rigorously analyzed using frameworks like:</p><ul><li><strong>PAC Learning (Probably Approximately Correct)</strong></li><li><strong>VC Dimension (Vapnik–Chervonenkis)</strong></li><li><strong>Empirical Risk Minimization (ERM)</strong></li></ul><p><strong>What Is Machine Learning — Really?</strong></p><p>At its core, <strong>machine learning is the science of making computers learn patterns from data without being explicitly programmed</strong>. It’s the evolution of computer science into a data-driven art form, guided by probability, optimization, and inference.</p><p>Unlike traditional software, where every scenario is hand-coded by humans, ML systems build their own logic by learning from examples. This transition — from explicit rules to implicit learning — mirrors how humans evolve from experience.</p><p><strong>The Three Learning Paradigms: Supervised, Unsupervised, and Reinforcement Learning</strong></p><p><strong>1. Supervised Learning — Learning from Labeled Data</strong></p><p>Imagine showing a child a thousand pictures of dogs and cats and telling them which is which. Eventually, they learn to identify them on their own. That’s supervised learning.</p><figure><img alt="http://localhost:8888/notebooks/ml%20article%20%201%20codes.ipynb" src="https://cdn-images-1.medium.com/max/831/1*fLJ5weLs5A8YGMbf_a9H9w.png" /></figure><ul><li><strong>Use cases</strong>: Email spam detection, sentiment analysis, medical diagnosis.</li><li><strong>Challenge</strong>: Requires labeled data, which is expensive and time-consuming to produce.</li></ul><figure><img alt="https://www.vproexpert.com/what-is-the-difference-between-supervised-and-unsupervised-machine-learning/" src="https://cdn-images-1.medium.com/max/878/1*Bs7z1ngeCH1uao1lgQg3DQ.png" /></figure><p><strong>2. Unsupervised Learning — Finding Hidden Patterns</strong></p><p>Here, the system isn’t told what to look for — it must find structure on its own. Think of it as giving an algorithm a box of puzzle pieces with no picture to guide it.</p><ul><li><strong>Use cases</strong>: Customer segmentation, topic modeling, anomaly detection.</li><li><strong>Unique strength</strong>: Reveals unexpected insights.</li></ul><p><strong>3. Reinforcement Learning (RL) — Learning by Interaction</strong></p><p>Inspired by behavioural psychology, RL involves agents learning to make decisions by interacting with an environment and receiving feedback (rewards or penalties).</p><ul><li><strong>Use cases</strong>: Robotics, game playing (e.g., AlphaGo), dynamic pricing.</li><li><strong>Special insight</strong>: RL optimizes <strong>long-term cumulative rewards</strong>, not just immediate outcomes.</li></ul><p><strong>Learning Is Not Guessing: The Mathematics Behind ML</strong></p><p>One myth worth busting: machine learning is not just sophisticated curve fitting. It’s underpinned by powerful theoretical guarantees.</p><p>Concepts like:</p><ul><li><strong>PAC Learning (Probably Approximately Correct)</strong>: Provides bounds on how well an algorithm will generalize to unseen data.</li><li><strong>VC Dimension</strong>: Measures the capacity of a model to fit various functions — essentially controlling overfitting.</li><li><strong>Bias-Variance Tradeoff</strong>: Balancing the model’s ability to generalize (bias) and capture complexity (variance).</li></ul><p>These ideas, explored in <em>Foundations of Machine Learning</em> by Mohri et al., are rarely discussed in popular blogs but are crucial for a deep understanding.</p><figure><img alt="https://www.vproexpert.com/what-is-the-difference-between-supervised-and-unsupervised-machine-learning/" src="https://cdn-images-1.medium.com/max/858/1*J5dq8Dcq2ANlItYLERL-uA.png" /></figure><p><strong>Data as the True Currency</strong></p><p>Data is not just the fuel — it’s the DNA of a machine learning system. Each row (called a <strong>sample</strong>) and each column (<strong>feature</strong>) carries semantic weight.</p><p>Good models don’t emerge from raw data — they are built on <strong>well-crafted features</strong>, a process called <strong>feature engineering</strong>. Think of this as teaching the model what to look at.</p><p><strong>Where Things Are Going: From Learning to Teaching</strong></p><p>The future of machine learning lies in systems that can <strong>teach themselves better representations</strong> — this is the idea behind <strong>self-supervised learning</strong>, which powers models like GPT and BERT.</p><p>Moreover, <strong>meta-learning</strong> (learning to learn) and <strong>causal inference</strong> (understanding cause-effect) are redefining what’s possible — moving beyond correlation toward reasoning.</p><p><strong>Final Thoughts: ML as a Philosophy of Problem Solving</strong></p><p>Machine learning is more than a toolkit — it’s a <strong>philosophy of designing systems that improve with experience</strong>. Whether you’re optimizing a delivery route or decoding the human genome, ML offers a framework that adapts, learns, and evolves.</p><p>But to truly unlock its power, we must go beyond plug-and-play, embracing the mathematical, ethical, and conceptual depths that underpin it.</p><p>In the words of Arthur Samuel, who coined the term “machine learning” in 1959:</p><figure><img alt="https://www.spiceworks.com/tech/artificial-intelligence/articles/what-is-ml/" src="https://cdn-images-1.medium.com/max/720/1*_zrZXwM7PWMLhjp7PlMhog.png" /></figure><p><strong>References:</strong></p><ol><li>Mohri , M., Rostamizadeh, A., &amp; Talwalkar, A. (2018). <em>Foundations of Machine Learning</em>. MIT Press.</li><li>Müller, A. C., &amp; Guido, S. (2017). <em>Introduction to Machine Learning with Python</em>. O’Reilly Media.</li><li>Vapnik ,V. N. (1998). <em>Statistical Learning Theory</em>. Wiley-Interscience.</li></ol><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=3b76a2c35e08" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>