<?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 Pushkar Raj on Medium]]></title>
        <description><![CDATA[Stories by Pushkar Raj on Medium]]></description>
        <link>https://medium.com/@rustydigg918?source=rss-22d21a76d6c7------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/2*yVyN6PHY_nUTEoJ7mTMgFQ.jpeg</url>
            <title>Stories by Pushkar Raj on Medium</title>
            <link>https://medium.com/@rustydigg918?source=rss-22d21a76d6c7------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Wed, 18 Nov 2026 20:36:54 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@rustydigg918/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[Google’s Meena: Why it’s a breakthrough]]></title>
            <link>https://medium.com/analytics-vidhya/googles-meena-why-it-s-a-breakthrough-19238b1d5023?source=rss-22d21a76d6c7------2</link>
            <guid isPermaLink="false">https://medium.com/p/19238b1d5023</guid>
            <category><![CDATA[googles-meena]]></category>
            <category><![CDATA[chatterbot]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[gｏｏｇｌｅ]]></category>
            <category><![CDATA[naturallanguageprocessing]]></category>
            <dc:creator><![CDATA[Pushkar Raj]]></dc:creator>
            <pubDate>Sun, 02 Feb 2020 17:47:34 GMT</pubDate>
            <atom:updated>2020-02-04T08:13:58.537Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*KYNyVs4Q9P_cz9I7XzABag.jpeg" /><figcaption>A shot from the movie <a href="https://www.imdb.com/title/tt1798709/">Her</a></figcaption></figure><p>Just like me if you too were amazed by Spike Jonze’s directorial ‘Her’ where the protagonist Theodore (played by <a href="https://www.google.com/search?client=firefox-b-d&amp;sxsrf=ACYBGNRSjKKTZUunbY-Mg4qI3u8mDVFPjQ:1580656994486&amp;q=Joaquin+Phoenix&amp;stick=H4sIAAAAAAAAAONgFuLSz9U3yCk0S0syVeIEsQ0tUpIstHh888syU0PyHZNL8osWsfJ75ScWlmbmKQRk5KfmZVbsYGUEADusTqI8AAAA&amp;sa=X&amp;ved=2ahUKEwi06auVlrPnAhVGyzgGHfVaCDoQxA0wNXoECB4QBQ">Joaquin Phoenix</a>) falls in love with an AI Samantha(played by <a href="https://www.google.com/search?client=firefox-b-d&amp;sxsrf=ACYBGNRSjKKTZUunbY-Mg4qI3u8mDVFPjQ:1580656994486&amp;q=Scarlett+Johansson&amp;stick=H4sIAAAAAAAAAONgFuLSz9U3yCk0S0syVeIEsY0MDU1KtHh888syU0PyHZNL8osWsQoFJycW5aSWlCh45Wck5hUX5-ftYGUEALTjhPo_AAAA&amp;sa=X&amp;ved=2ahUKEwi06auVlrPnAhVGyzgGHfVaCDoQxA0wNXoECB4QCQ">Scarlett Johansson</a>). Actually, this movie was one of the reasons which hypnotized me to chose Computer Science(especially Machine Learning &amp; Artificial Intelligence) as my professional career and so whenever some breakthrough like this is made in the field, it feels like a personal gain to me, that’s why i’ve decided to write this blog.</p><h3>Overview</h3><p><strong><em>What is Meena?</em></strong></p><p><strong><em>What is its history?</em></strong></p><p><strong><em>What will be its impact?</em></strong></p><p><strong><em>Why it’s a breakthrough in right direction for multi-turn chat bots?</em></strong></p><p><strong><em>what to expect now?</em></strong></p><blockquote><strong><em>What is Meena?</em></strong></blockquote><p>Meena is a chatbot <a href="https://ai.googleblog.com/2020/01/towards-conversational-agent-that-can.html">released by Google</a> recently claiming it to be the best voice assistant at the moment. We already have some very popular chatbots like Apple’s Siri, Google’s Google Assistant, Amazon’s Alexa, Microsoft’s Cortana, etc. But they aren’t good enough as the quest of all the virtual assistant AI is to match human-like experience which seems to be very near after looking at the results of Meena.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*G4tHSRLLsSRrlLsRxD7iCA.png" /></figure><p><strong>Now, what is this Sensibleness and Specificity Average (SSA)?</strong></p><p>It’s a human evaluation metric that decides how humanly(Sensibly and Specifically) do you answer. This metric is designed by Google which argues that SSA captures basic yet important attributes for natural conversations.</p><blockquote><strong><em>What is its history?</em></strong></blockquote><p>Chatbots have been here for more than fifteen years now, and tech giants like Google, Facebook, Amazon, Microsoft, IBM, etc are betting big on this, the availability of chatbot frameworks like <a href="https://dialogflow.com/">DialogueFlow</a> from Google, <a href="https://www.ibm.com/watson/how-to-build-a-chatbot">Watson</a> from IBM’s, <a href="https://dev.botframework.com/">Bot Framework</a> from Microsoft, <a href="https://rasa.com/">RASA</a>(an open-source framework) seduces developers and enthusiasts to have their eye on it constantly in the name of achieving the technological singularity.</p><p><strong><em>ELIZA </em></strong>is considered to be the first chatbot in the history of Computer Science which was developed by <em>Joseph Weizenbaum</em> at MIT. ELIZA operates by recognizing sentences and keywords from its provided keywords to reproduce a response using those keywords from pre-programmed responses. Like, Someone says ‘I love my Company’ then ELIZA would pick up the word ‘Company’, and reply by asking an open-ended question ‘Wow! Tell me something about your company’. Which felt like an Auto-Chatter but it was just a programmed response and nothing more but that was the inception of chatbots and our quest for humans like chatter which was later followed by <em>ALICE(a </em>natural language processing powered bot) and Steve Worswick’s <em>MITSUKU</em></p><p>These days even you may go on to create your own chatbot using these above-mentioned frameworks with a very minimal effort. I highly encourage you to go on and try it (especially if you’re not from a Computer science background.)</p><blockquote><strong><em>What will be its impact?</em></strong></blockquote><p>Since <strong>Meena</strong> model has 2.6 billion parameters and is trained on 341 GB of text, filtered from public domain social media conversations for thirty(30) days on a TPU v3 pod (2,048 TPU cores). Compared to an existing state-of-the-art generative model, <a href="https://openai.com/blog/better-language-models/">OpenAI GPT-2</a>, Meena has 1.7x greater model capacity and was trained on 8.5x more data.</p><p>More data means better results.</p><blockquote><strong><em>Why it’s a breakthrough in the right direction for multi-turn chatbots?</em></strong></blockquote><p>Although Google, Amazon and Microsoft’s assistant are hugely successful they are miles away from Samantha from the movie “HER” :-)</p><p><em>Another reason why our expectations are so high because Google has recently achieved the Quantum Boost, So it’ll be easier, faster and far more efficient for them to achieve a human-like chatbot than any of its adversaries.</em></p><blockquote><strong><em>What to expect now?</em></strong></blockquote><p>Although Google claims it to be the best but at its heart lies the <a href="https://en.wikipedia.org/wiki/Seq2seq">seq2seq</a> architecture which isn’t something completely revolutionary and is common in the community but since Google has quantum power which allows them to feed Meena a huge data to achieve better results which means we’ve come a long way but clearly not there yet.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=19238b1d5023" width="1" height="1" alt=""><hr><p><a href="https://medium.com/analytics-vidhya/googles-meena-why-it-s-a-breakthrough-19238b1d5023">Google’s Meena: Why it’s a breakthrough</a> was originally published in <a href="https://medium.com/analytics-vidhya">Analytics Vidhya</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[K-Nearest Neighbors: All you need to know]]></title>
            <link>https://medium.com/analytics-vidhya/k-nearest-neighbors-all-you-need-to-know-1333eb5f0ed0?source=rss-22d21a76d6c7------2</link>
            <guid isPermaLink="false">https://medium.com/p/1333eb5f0ed0</guid>
            <category><![CDATA[k-nearest-neighbours]]></category>
            <category><![CDATA[data-science]]></category>
            <category><![CDATA[supervised-learning]]></category>
            <category><![CDATA[machine-learning]]></category>
            <category><![CDATA[lazy-learner]]></category>
            <dc:creator><![CDATA[Pushkar Raj]]></dc:creator>
            <pubDate>Mon, 11 Nov 2019 21:02:30 GMT</pubDate>
            <atom:updated>2019-11-18T17:04:00.636Z</atom:updated>
            <content:encoded><![CDATA[<p>Have never been a fan of showing off my knowledge by blogging rather than writing papers or even a book but felt rusty explaining K- Nearest Neighbors(K-NN) to one of my colleagues earlier today. Hence in this post i’m explaining everything you need to know about K-NN .</p><blockquote>You are the average of those five humans living around you.</blockquote><p><strong><em>Table of Contents:</em></strong></p><ol><li>Introduction</li><li>How it works</li><li>What is this <strong>‘K</strong>’ in K-NN(also called hyper parameter tuning in this case)</li><li>Creating Movies Recommender System using K-NN</li><li>Where to go from here</li></ol><figure><img alt="Nearest Neighbor" src="https://cdn-images-1.medium.com/max/1024/1*HG85cPnoo214nIx2qIh_Eg.png" /><figcaption>My neighbor does this so i’m doing it as well.</figcaption></figure><h3>{1} Introduction</h3><p>Despite being considered as one of the simplest machine learning algorithms(maybe because of easy to relate examples around us), K-NN has proved to be insanely useful while solving the real world problems.</p><h4>When to use:</h4><p>It is a supervised machine learning algorithm(output labels are provided) which <strong>solves both </strong><a href="https://www.cs.princeton.edu/~schapire/talks/picasso-minicourse.pdf"><strong>Classification</strong></a><strong> and </strong><a href="https://www.kdnuggets.com/2017/02/regression-analysis-primer.html"><strong>Regression</strong></a><strong> problems</strong>. Although it is widely used for the classification problems, it’s effective at regression problems as well but the usage is minuscule.</p><h3>{2} How does it work</h3><p>Let’s understand K-NN with a simple example where we have to find the value of the empty circle based on distance to its nearest neighbors(Mind you that i’ve considered Minkowski distance of degree ‘2’ here). Since in the red boundary we could see that there is one negative values nearest to our empty circle so we assign the value of the empty circle as negative. Obvious, right??</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/462/1*nzzIRMVUkve2GAR2kBksnA.png" /><figcaption>My Nearest-Friend</figcaption></figure><p><strong>NOTE: </strong><em>Mind you that ideally, you should not consider an even value of K while executing K-NN, as deciding the outcome on 50–50 split might create an ambiguity.</em></p><blockquote>Different ways to calculate the distance</blockquote><blockquote><a href="https://en.wikipedia.org/wiki/Minkowski_distance">Minkowski distance</a>( Euclidean, Manhattan etc)</blockquote><blockquote><a href="https://en.wikipedia.org/wiki/Chebyshev_distance">Chebyshev distance</a></blockquote><blockquote><a href="https://en.wikipedia.org/wiki/Cosine_similarity">Cosine Similarity</a></blockquote><blockquote><a href="https://en.wikipedia.org/wiki/Hamming_distance">Hamming distance</a></blockquote><p>Now, let&#39;s look at the picture below where we need to find the value of the empty circle on the basis of its neighbors, but we’ve just encountered a dilemma of choice.</p><figure><img alt="Change in value by change in value of k" src="https://cdn-images-1.medium.com/max/591/1*gn48DrMdZLTulYSBAuxLCw.png" /></figure><blockquote>The inner purplish boundary points the value to be negative while the outer yellow boundary says it’s positive. That’s where the concept of <strong>K </strong>in K-NN comes along.</blockquote><h3>{3} What is this <strong>‘K</strong>’ in K-NN</h3><p>Let’s consider a real time example to find the value of K in it. That’s how our dataset looks like and we’re going to find the value of K for this problem.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/782/1*94LDm48zOSedDrV3L3Rffg.png" /></figure><p>Now, There are two popular ways to find the value of K in K-NN</p><blockquote>(a) Elbow Curve</blockquote><blockquote>(b) Grid Search</blockquote><p>Let’s look at Elbow Curve to find the value of K</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/680/1*uOInJpQgxt6cnu8hjt-61w.png" /><figcaption>Elbow Curve</figcaption></figure><p>So, our elbow curve says the value of K = 8.</p><p>So, once we find out the value of K, the problem is solved. Now all we need to do is predict the value and look out for our accuracy measures. Let’s have a look at it.(you may find the Notebook <a href="https://jovian.ml/rustydigg918/24418db5df5e4224a17b7e1331d3703f">here</a>)</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/595/1*z0ksQFSsrDKyi9XsPnXfkg.png" /><figcaption>Accuracy Score</figcaption></figure><h3>{4} Movies Recommender System using K-NN</h3><p>Recommender systems can be loosely broken down into three categories:</p><blockquote>&gt;&gt;&gt; <strong>Content Based Systems:</strong> It utilizes a series of discrete characteristics of an item in order to recommend additional items with similar properties.</blockquote><blockquote>&gt;&gt;&gt; <strong>Collaborative Filtering Systems: </strong>It builds a model from a user’s past behaviors (items previously purchased or selected and/or numerical ratings given to those items) as well as similar decisions made by other users. This model is then used to predict items (or ratings for items) that the user may have an interest in.)</blockquote><blockquote>&gt;&gt;&gt; <strong>Hybrid Systems: </strong>It combines the previous two approaches. Most businesses prefer to use hybrid approach in their production recommender systems.</blockquote><blockquote>Step 1: Importing necessary libraries</blockquote><blockquote># Import necessary libraries<br>import matplotlib.pyplot as plt<br>import pandas as pd<br>import numpy as np<br>import seaborn as sns<br>%matplotlib inline<br>from IPython.display import Image<br># utils import<br>from fuzzywuzzy import fuzz</blockquote><blockquote>Step 2: Reading the dataset</blockquote><blockquote># read data<br>df_movies = pd.read_csv(<br> ‘<a href="http://khansingh.xyz/Datasets/movies.csv&#39;">http://khansingh.xyz/Datasets/movies.csv&#39;</a>,<br> usecols=[‘movieId’, ‘title’],<br> dtype={‘movieId’: ‘int32’, ‘title’: ‘str’})</blockquote><blockquote>df_ratings = pd.read_csv(<br> ‘<a href="http://khansingh.xyz/Datasets/ratings.csv&#39;">http://khansingh.xyz/Datasets/ratings.csv&#39;</a>,<br> usecols=[‘userId’, ‘movieId’, ‘rating’],<br> dtype={‘userId’: ‘int32’, ‘movieId’: ‘int32’, ‘rating’: ‘float32’})</blockquote><p>Note: Although KNN does not make any assumptions on the underlying data distribution but it relies on item feature similarity. When KNN makes inference about a movie, it will calculate the “distance” between the target movie and every other movie in its database, then it ranks its distances and returns the top K nearest neighbor movies as the most similar movie recommendations.</p><blockquote>Step 3: Data Manipulation</blockquote><blockquote>from scipy.sparse import csr_matrix<br># pivot ratings into movie features<br>df_movie_features = df_ratings.pivot(<br> index=’movieId’,<br> columns=’userId’,<br> values=’rating’<br>).fillna(0)</blockquote><blockquote># create mapper from movie title to index<br>movie_to_idx = {<br> movie: i for i, movie in <br> enumerate(list(df_movies.set_index(‘movieId’).loc[df_movie_features.index].title))<br>}</blockquote><blockquote># convert dataframe of movie features to scipy sparse matrix<br>mat_movie_features = csr_matrix(df_movie_features.values)</blockquote><p>Now, Have a look at our derived dataset which is sparse for the obvious reason.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/734/1*pDgoV8e-lTzaj1T4FoDpIA.png" /><figcaption>Sparse table</figcaption></figure><p>Let’s enhance the visibility of the table</p><blockquote>df_movie_features_name = df_complete.pivot_table(<br> index=’userId’,<br> columns=’title’,<br> values=’rating’<br>).fillna(0)</blockquote><blockquote>Step 4: Data Visualization</blockquote><blockquote>ax = df_movie_cnt \<br> .sort_values(‘movie_cnt’, ascending=False) \<br> .reset_index(drop=True) \<br> .plot(<br> figsize=(12, 8),<br> title=’Rating Frequency of All Movies’,<br> fontsize=12<br> )<br>ax.set_xlabel(“movie Id”)<br>ax.set_ylabel(“number of ratings”)</blockquote><p>Now, Let’s look at this graph</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/847/1*icF-9puKQHIWwI-wCz-GUg.png" /><figcaption>Taking idea of our dataset</figcaption></figure><p>As we can see it through the data that vast majority of people aren’t interested in rating a movie so, we could consider only top 40% of our data.</p><blockquote># filter data<br>ratings_thres = 50<br>active_users = list(set(df_users_cnt.query(‘movie_cnt &gt;= <a href="http://twitter.com/ratings_thres">@ratings_thres</a>’).index))<br>df_ratings_drop_users = df_ratings_drop_movies[df_ratings_drop_movies.userId.isin(active_users)]<br>print(‘shape of original ratings data: ‘, df_ratings.shape)<br>print(‘shape of ratings data after dropping both unpopular movies and inactive users: ‘, df_ratings_drop_users.shape)</blockquote><blockquote>Step 5: Now, one of the most crucial part of creating the model, choosing the appropriate distance measure to calculate the value of K</blockquote><p>Notice that we have got a pretty high dimensional dataset to deal with While it was required to do what we did, but high dimensionality is not good for KNN.<br>Also, Euclidean distance is unhelpful in high dimensions because all vectors are almost equidistant to the search query vector (target movie’s features). Instead, we will use cosine similarity for nearest neighbor search.</p><p>The cosine similarity is advantageous because even if the two similar vectors or documents are far apart by the Euclidean distance (due to the size of the document), chances are they may still be oriented closer together. <br>The smaller the angle, higher the cosine similarity.</p><p>When plotted on a multi-dimensional space, where each dimension corresponds to a feature in the feature vector, the cosine similarity captures the orientation (the angle) of the feature and not the magnitude. If you want the magnitude, compute the Euclidean distance instead.</p><p>There is also another popular approach to handle nearest neighbor search in <br>high dimensional data, locality sensitive hashing.</p><blockquote>from sklearn.neighbors import NearestNeighbors<br>model_knn = NearestNeighbors(metric=’cosine’, algorithm=’brute’, n_neighbors=20, n_jobs=-1)</blockquote><p>Fitting the model:</p><blockquote>model_knn.fit(mat_movie_features)</blockquote><p><strong>Let’s make the prediction readable</strong></p><pre>def fuzzy_matching(mapper, fav_movie, verbose=True):<br> “””<br> return the closest match via fuzzy ratio. If no match found, return None<br> <br> Parameters<br> — — — — — <br> mapper: dict, map movie title name to index of the movie in data</pre><pre>fav_movie: str, name of user input movie<br> <br> verbose: bool, print log if True</pre><pre>Return<br> — — — <br> index of the closest match<br> “””<br> match_tuple = []<br> # get match<br> for title, idx in mapper.items():<br> ratio = fuzz.ratio(title.lower(), fav_movie.lower())<br> if ratio &gt;= 60:<br> match_tuple.append((title, idx, ratio))<br> # sort<br> match_tuple = sorted(match_tuple, key=lambda x: x[2])[::-1]<br> if not match_tuple:<br> print(‘Oops! No match is found’)<br> return<br> if verbose:<br> print(‘Found possible matches in our database: {0}\n’.format([x[0] for x in match_tuple]))<br> return match_tuple[0][1]</pre><pre>def make_recommendation(model_knn, data, mapper, fav_movie, n_recommendations):<br> “””<br> return top n similar movie recommendations based on user’s input movie</pre><pre>Parameters<br> — — — — — <br> model_knn: sklearn model, knn model</pre><pre>data: movie-user matrix</pre><pre>mapper: dict, map movie title name to index of the movie in data</pre><pre>fav_movie: str, name of user input movie</pre><pre>n_recommendations: int, top n recommendations</pre><pre>Return<br> — — — <br> list of top n similar movie recommendations<br> “””<br> # fit<br> model_knn.fit(data)<br> # get input movie index<br> print(‘You have input movie:’, fav_movie)<br> idx = fuzzy_matching(mapper, fav_movie, verbose=True)<br> # inference<br> print(‘Recommendation system start to make inference’)<br> print(‘……\n’)<br> distances, indices = model_knn.kneighbors(data[idx], n_neighbors=n_recommendations+1)<br> # get list of raw idx of recommendations<br> raw_recommends = \<br> sorted(list(zip(indices.squeeze().tolist(), distances.squeeze().tolist())), key=lambda x: x[1])[:0:-1]<br> # get reverse mapper<br> reverse_mapper = {v: k for k, v in mapper.items()}<br> # print recommendations<br> print(‘Recommendations for {}:’.format(fav_movie))<br> for i, (idx, dist) in enumerate(raw_recommends):<br> print(‘{0}: {1}, with distance of {2}’.format(i+1, reverse_mapper[idx], dist))</pre><blockquote>Step 6: Let’s predict the movie:</blockquote><blockquote>my_favorite = ‘Pulp Fiction (1994)’</blockquote><blockquote>make_recommendation(<br> model_knn=model_knn,<br> data=mat_movie_features,<br> fav_movie=my_favorite,<br> mapper=movie_to_idx,<br> n_recommendations=10)</blockquote><figure><img alt="" src="https://cdn-images-1.medium.com/max/1014/1*1kzZgNMJ9hsd-29bFjkrWg.png" /><figcaption>The recommendation</figcaption></figure><p>Although we’ve successfully recommended the movie(Link to Notebook is <a href="https://github.com/rustydigg918/movie_recommender/">here</a> but in the mean time we’ve just effectively identified that there are two shortcomings in item based collaborative filtering:</p><blockquote>1. Popularity Bias: recommender is prone to recommender popular items</blockquote><blockquote>2. Item Cold-Start Problem: recommender fails to recommend new or less-known items because items have either none or very little interactions imagine for a new movie with no or little ratings</blockquote><h3>{5} Where to go from here</h3><p>Improvements to the current system</p><p>This movie recommender could be further improvised using “Matrix Factorization”</p><p>In a real world setting, the vast majority of movies receive very few or <br>even no ratings at all by users. We are looking at an extremely sparse matrix <br>with more than 99% of entries are missing values. With such a sparse matrix, what ML algorithms can be trained and reliable to make inference? So, data scarcity problem has to be addressed to build a better model and Matrix Factorization is the solution.</p><p>And that will be covered in my next post.</p><p>References:</p><blockquote><em>Springer’s An Introduction to Statistical Learning</em></blockquote><blockquote>Mit opencourseware for machine learning</blockquote><blockquote>Introductory Applied <em>Machine Learning</em> (IAML) course at the University of Edinburgh</blockquote><blockquote>Analytics Vidhya, Medium, KDnuggets</blockquote><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=1333eb5f0ed0" width="1" height="1" alt=""><hr><p><a href="https://medium.com/analytics-vidhya/k-nearest-neighbors-all-you-need-to-know-1333eb5f0ed0">K-Nearest Neighbors: All you need to know</a> was originally published in <a href="https://medium.com/analytics-vidhya">Analytics Vidhya</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
    </channel>
</rss>