<?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 Shahar Cohen on Medium]]></title>
        <description><![CDATA[Stories by Shahar Cohen on Medium]]></description>
        <link>https://medium.com/@shahar.cohen?source=rss-ab3b8d6f6689------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*HYSa2rZ8Sw5Re5z5JZE3EA.jpeg</url>
            <title>Stories by Shahar Cohen on Medium</title>
            <link>https://medium.com/@shahar.cohen?source=rss-ab3b8d6f6689------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Sun, 22 Nov 2026 01:33:09 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@shahar.cohen/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[How to partner and work together in early startup stages]]></title>
            <link>https://medium.com/@shahar.cohen/how-to-partner-and-work-together-in-early-startup-stages-b28c73b19b2?source=rss-ab3b8d6f6689------2</link>
            <guid isPermaLink="false">https://medium.com/p/b28c73b19b2</guid>
            <category><![CDATA[partn]]></category>
            <category><![CDATA[cofounders]]></category>
            <category><![CDATA[startup]]></category>
            <dc:creator><![CDATA[Shahar Cohen]]></dc:creator>
            <pubDate>Fri, 11 Jan 2019 18:46:30 GMT</pubDate>
            <atom:updated>2019-01-11T18:46:30.657Z</atom:updated>
            <content:encoded><![CDATA[<p>During my career, I have co-founded two startups of my own and provided professional services to more than two dozen others. I have always been intrigued by unicorn entrepreneurs who found and lead their startups on their own. Quite early in my career, I learned that I cannot be one of them, and if you are a lonely unicorn, maybe this post is not for you. For me, at least, thinking about a new venture immediately leads me to consider partners. And good partners, as my uncle once said, are very hard to find:</p><blockquote>“If a good partner was so easy to find, then God would already have one”</blockquote><p>In this post, I share a synthesis of my experience and observations on partnerships in the form of eight tips on selecting the right partners and working with them during the early stages of new ventures. I hope it will shed some light on the soft parts of partnerships — one of the most important ingredients in successful enterprises.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*ddvR32oMhn3xpCfi" /><figcaption>Photo by <a href="https://unsplash.com/@rawpixel?utm_source=medium&amp;utm_medium=referral">rawpixel</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a></figcaption></figure><p><strong>1.</strong> <strong>My uncle is right, but get yourself partnered anyway</strong></p><p>Partnering around a venture is hard to maintain. Playing with strong, opinionated, emotional individuals who see the world through different goggles to yours — operating together with them in a stressful reality while being totally exposed in your weaknesses, sometimes criticized, and constantly expected to produce unique value — is a challenging starting point. Yet this is indeed the ground on which all venture partnerships are built. No wonder, therefore, that despite the need and great benefit of partnerships between entrepreneurs, many of them inevitably (and often quite quickly) break up.</p><p>But, raising a new venture is a long, stormy and challenging journey. Challenging in terms of the mental complexity that must be dealt with and the emotional load involved. It is a journey that takes entrepreneurs to their extremes. At times, you feel so down that you can hardly hold it anymore. But there are other times. Wonderful times, when you see the highest peaks and experience real self-fulfillment. On this mental and emotional roller coaster, it is hard to be alone. And indeed, it is well known that the success rate of ventures founded by single entrepreneurs is low.</p><p>There may be some unicorns that can grow successful ventures alone, and even a few that are better off alone. But I think that for most of us humans, founding a business venture alone is simply too hard. So, if you want to start up a new venture, go and search for the right partners.</p><p><strong>2.</strong> <strong>Finding the right partners: The thin line between similarity and diversity</strong></p><p>When it comes to starting up a business venture, the main reason for engaging with partners is to achieve success. Friendship and mutual enjoyment are important, but they are not the bottom line. Therefore, one should seek for business partners whose presence will maximize the probability of the business venture becoming a success.</p><p>Most of the time, it is easy for us to feel comfortable with people who are like us — who can easily understand us. The conversation is flowing; we get strong positive reinforcement of our thoughts. However, partners who are too much like us will not challenge us, will not support us in situations that we cannot cope with well enough ourselves, and they may aspire to a role that we had assumed would go to us. Therefore, diversity is a key consideration in maximizing the probability of success when working with partners.</p><p>The real tragedy of partnerships, however, is that maintaining them for long periods of time with those who are different from us is unfortunately very difficult. Building the most effective partnerships requires a combination of similarity and harmony on the one hand and diversity and synergy on the other. In general, I recommend examining similarity and diversity in three dimensions, which are virtually independent of each other:</p><p><strong>Values: Strive to achieve a reasonable match</strong></p><p>Your set of values is the set of fundamental beliefs that you have with respect to different aspects of reality. It is the compass that directs you in everything you do. Therefore, values are not the recommended dimension in which to search for diversity.</p><p>In general, partners are more likely to form an organization with a clear identity when they hold similar sets of values. If your partners’ vision of success is significantly different from yours, if the organizational culture that they imagine is different from yours, if they have a very different sense of life balance, if they are significantly more receptive to risk than you are (or significantly less), the chances are that over time some of you may find yourselves acting in direct contradiction to your values. You will likely feel frustrated and sometimes angry at your partners, without them ever understanding why.</p><p><strong>Professional background: Strive to achieve reasonable diversity</strong></p><p>The diversity of professional backgrounds between partners is a basis for synergy, given that the different backgrounds are relevant to the venture. There isn’t much point in partnering with someone who has the same professional background as you. You might find yourselves with highly overlapping areas of responsibility and frequent disagreements. Moreover, growing your new venture requires many different sets of skills, which are more likely to be present among partners with different professional backgrounds.</p><p><strong>Internal operating systems: Diversity is good, but it makes life more complex</strong></p><p>Your internal operating system is the very large set of methods and mechanisms that you apply during your daily life and the intuitive way that you respond to different situations. It is independent of your set of values and, clearly, of your professional background.</p><p>For example, for some individuals, brainstorming is an ideal way to achieve their objectives, for others it may be quite the opposite. Some capture reality in a top-down manner, others don’t understand their surroundings until they gather all the required details. Some people are at their best during noisy and emotional discussions, others strongly prefer sessions of cold logic, in which participants take turns to state their opinions. Our operating systems are complex mechanisms that evolve according to our basic biology. You cannot escape from your own operating system.</p><p>Partners who have similar operating systems will have more harmonic relationships, experience less frustration and are likely to enjoy carrying out their mutual work together for much longer. On the other hand, such partners will tend to challenge each other less and will have fewer ways to react to various, sometimes critical, situations. For examples, two partners who are both creative and idea-driven might find it hard to materialize their wonderful ideas. An idea-driven entrepreneur partnered with a detail- and operations-driven partner might feel less excited during mutual brainstorming, but they will enjoy a totally different way of getting things done.</p><p>Working with partners whose systems are completely different from yours is, in many cases, like working with aliens. You need to know them well, be sensitive to their needs and sometimes behave in a way you normally wouldn’t in order to serve their needs. You must be ready for any frustrations that might arise. You will have to invest a lot of energy in them, in order to resolve them.</p><p>One of my first observations on different operating systems occurred with my partner in Optimove.com. As an individual, I am highly driven towards new ideas. I play a lot with new ideas. I am excited by new ideas and have a strong urge to share them. I am also aware that ideas are just ideas. They might be helpful and brilliant, but they might also be revealed as nonsense. At that time, whenever I came up with an idea, I would share it excitedly with my partner. Feeling my excitement, my partner felt that I was eager to materialize my ideas, which would significantly change the course of our business. That wasn’t the case. I just wanted to discuss ideas, but my partner felt that instead of stabilizing reality (which was shaky at the time even without my new ideas) I was creating total chaos. In self-defense, he used to reject my ideas without even listening to them. That, of course, led to very bad feelings from my end. We very quickly recognized the pattern and reached an understanding that worked beautifully for both of us. Every time I wanted to share a new idea, I asked his permission first. I used to ask him if it was okay to shake things up for a moment. Forewarned that I was knowing that I was simply putting forward ideas and not actual work plans, my partner had no problem participating in the brainstorm.</p><p><strong>3.</strong> <strong>Be direct and open right from day one</strong></p><p>Where in other relationships you have the privilege to progress one step at a time, and it’s okay not to be sure that you are a perfect match, in a business venture you don’t have time. The longer you wait, the more invested you become in the new venture, the more committed you become (to your partners, customers, investors, employees) and the more you come to depend on each other. Therefore, it is vital that you discover any mismatches as early in the process as possible. One way of doing this is to talk a lot. To raise all your concerns. To align expectations in all the relevant aspects that you are involved in. To examine the monetary needs of each partner. To sketch roles and responsibilities, life balance, potential risks, decision-making mechanisms, and any other important aspect of the mutual venture.</p><p>Although romance is not the purpose of partnerships, I am sure that once partnering you will soon learn that romance is there in those early stages. And when romance is around, there is a tendency to want to preserve it, while repressing subjects that have flammable potential. Don’t do that. Talk about everything and ask your partners to do the same. My rule of thumb is that the pains that you feel in the early stages of the partnership will hurt you far more in the later stages.</p><p>In those early stages, don’t take anything for granted. If you didn’t ask your partners about their attitude towards something, don’t make assumptions. Approach them and raise the issue. I especially advise you to describe your aspirations for the new venture. A known syndrome is the <em>experience-versus-plans</em> discrepancy. In many cases, entrepreneurs have a very strong drive, which is way wider than their actual experience. It is not necessarily that your technological partner will seek to fill the technological role in the venture, or that your experienced sales manager partner will not aspire to be involved with the product, and so on. Do not assume anything. Ask. Authentically describe your aspirations and carefully examine whether they align with the plans of your partners.</p><p><strong>4.</strong> <strong>Stress yourselves as early as possible</strong></p><p>In many ventures, the load in the early days is relatively light. You don’t owe anything to anyone. There’s just you, the founding partners. And, as with romances, many new ventures take time to get up to speed. I warmly suggest you fight that slow starting pace. Challenge yourselves as early as possible. Decide on tasks, even if you still don’t know what you are trying to achieve. The reason it’s important to speed up quickly is not only the speed of the venture. Even more important than that is the actual value of experiencing some (healthy) stress as early as possible. To be able to recognize gaps while the dependencies are still weak. While you can still work together to shape your mutual work patterns, or even break up now (if you don’t think that things can be fixed). Although conversation (see Section 3 above), honesty and openness are important, real experience will tell you more.</p><p>Ask yourself how you felt being away from your comfort zone. Are you pleased with the way you operated? Are you pleased with the way your new partners operated? Use these early experiences for retrospective and mutual development.</p><p><strong>5.</strong> <strong>The balance between your compromises and expectations from other partners</strong></p><p>Partnerships are complex and require a lot of compromises. You must be willing to compromise. But compromising all the time is not healthy for the venture. Do not compromise all the time. Do not compromise over things that strongly conflict with your values. Do not agree to always do things in a way that is not right for you. Don’t compromise if you don’t think you will be able to live with the consequences in the long run.</p><p>Striking a balance between helping out your partners and making sure they put in the reciprocal effort, can be hard. Partly because of the romance you’re experiencing, it’s easier to please your partners and keep them sweet. However, as much as I believe a partnership cannot succeed without some compromises, I equally believe a partnership cannot succeed with too many compromises.</p><p><strong>6.</strong> <strong>Decide on roles and responsibilities as early as possible and work together under them</strong></p><p>In the early stages of a business venture, many partners work together with a lot of overlap. There is nothing wrong with that. It maintains the romance (which, despite what I wrote above, I think is a very strong asset in the early stages) and it helps your partners know each other better. But, very quickly, this overlap becomes disastrous. Requiring all the partners for every decision is extremely inefficient. And worse than that, doing everything together, multiplies the magnitude of disagreements and frustration.</p><p>Very early in the life cycle of the new venture, agree on the role and responsibilities of each partner. Avoid fuzzy borders as much as possible. It is important that every partner has the authority to decide in their own area of responsibility. It is important to share. It is important to reflect. It is important to hear more opinions, but a partner who cannot make decisions will not stay in the partnership for long.</p><p><strong>7.</strong> <strong>Be prepared, at a certain time, to accept the venture as an independent entity that is more important than you are</strong></p><p>Partnering and joining the venture is an important personal decision. It is part of your way to self-fulfillment — a very personal and egoistic move. However, when a business venture evolves, it becomes an entity in its own right. It has its own needs and its own dynamics, and this is the entity that you created. Playing a real partner’s role means that you will often find yourself investing in activities that you never thought you would; you will have to live with decisions that you do not agree with, but which affect your personal life dramatically. In the early stages of the venture, make sure that all the partners understand the huge responsibility they are taking on and its potential consequences.</p><p>A very common test in the very early stages of startups is how willing you are to quit your job in order to devote yourself to the new venture. Quitting a job at a very early stage implies huge personal risk, which significantly affects the personal lives of the partners. I do not necessarily recommend quitting your jobs as a default, but as the venture progresses, the time will inevitably come when you must be fully committed and refusing to do so is an indication of a partner’s unwillingness to pay a personal price for the venture.</p><p><strong>8.</strong> <strong>Write a founders’ agreement right from the start</strong></p><p>A founders’ agreement is a legal contract between the founders. It covers as many of their mutual agreements as possible. Writing a good founder’s agreement is not easy; it requires the services of a professional. You will have to think of potential scenarios — some of them will be negative and even depressing — and agree together on how you should act in such circumstances.</p><p>You will be writing the founders’ agreement when practically none of these scenarios will seem to be relevant. Do not be tempted to treat your founders’ agreement as anything less than the most important agreement that you will ever sign. Business ventures tend to change unexpectedly, and a scenario that seemed impossible just a few months ago can easily become reality. When circumstances change, the partners’ agendas may no longer align. Some rights that might seem trivial to you may seem outrageous to your partners, and vice versa. When strong disagreements arise, and mainly when they imply that some of the partners will quit their position in the company, the only stable ground you can rely on is your founders’ agreement.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=b28c73b19b2" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Evaluating the Error of Regression Models: Some Real-Life Challenges and Practical Tips]]></title>
            <link>https://medium.com/yellowblog/evaluating-the-error-of-regression-models-some-real-life-challenges-and-practical-tips-817b921a9a62?source=rss-ab3b8d6f6689------2</link>
            <guid isPermaLink="false">https://medium.com/p/817b921a9a62</guid>
            <category><![CDATA[regression]]></category>
            <category><![CDATA[rmse]]></category>
            <category><![CDATA[aem]]></category>
            <category><![CDATA[data-science]]></category>
            <category><![CDATA[mape]]></category>
            <dc:creator><![CDATA[Shahar Cohen]]></dc:creator>
            <pubDate>Mon, 02 Oct 2017 11:25:28 GMT</pubDate>
            <atom:updated>2017-10-02T11:25:28.602Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/450/1*iyZroKEGO_76KWpn955asg.png" /></figure><p>In regression, we predict (or estimate) the numerical value of an unknown quantity, denoted by <strong><em>y</em></strong>, as a function of some explaining features. The difference between the prediction and the actual value is the error. It is a numeric random variable, that might depend on the explaining features.</p><p>There are few common ways for estimating the performance and comparing several competing regression modes, which differ in their ways to measure error. Probably the most common measures are: Root Mean Square Error (RMSE), Mean Absolute Error (MAE) and Mean Absolute Percentage Error (MAPE). All these 3 measures are straight forward to compute and understand, however applying the right measure in real-life use cases is far from being trivial, and in most times, none of these measures cannot tell the entire story. In this post I describe some of the difficulties of evaluating the error of regression models, and suggest some practical tips on how to tackle these difficulties.</p><p><strong>The main limitations of RMSE, MAE and MAPE</strong></p><p>Here are two important limitations of RMSE, MAE and MAPE:</p><ol><li>Each of these measures is merely an average (or the square root of that average in the case of RMSE) of realizations of the test errors. The error is a numeric random variable and you cannot grasp the entire behavior of a random variable with a single aggregation of observations. The error is just a random variable, it is often a highly skewed random variable. When we predict skewed outcomes (like prices, incomes, user engagement, item sales and many many more) most probably that the error will also be skewed, meaning that in the majority of the cases the error is very small, but there are relatively rare examples that might have extremely large errors. When the error is highly skewed, the average often says nothing.</li><li>It might be the other side of the same coin, but it provides a different perspective. In average, all the observations equally contribute to the total error, regardless of the actual value of <strong><em>y</em></strong>. But error means something different, when associated to different values of <strong><em>y</em></strong>. For small values of <strong><em>y</em></strong> (e.g., near zero), we can accept relatively high percentage error, as long as the absolute error is small. For large values of <strong><em>y</em></strong> we may accept larger absolute errors, but not high percentage errors.</li></ol><p><strong>An Illustrative Example</strong></p><p>The following plot describes the predictions of a simple regression tree, on a common regression task (in which the objective is to predict sales prices).</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/668/1*vDMAw-EiiSwTI8KprXSJnA.png" /></figure><p>Each sale in the plot is a single point, where the X-axis describe the actual price, and the y axis describes the prediction by the regression tree. The tree has 6 leaves that separates sales into price ranges almost perfectly. Since the major part of the sales is associated to very low prices, and due to the scale of the plot, the first 3 ranges cannot be recognized. Such sales ranges may be found, for example, in online advertisement applications.</p><p>Since the model is a regression tree, a single value (leaf average) is associated with each price level, but the range of the actual prices in each level is different. Where in the low-price sales the range is limited, the highest range describe sale with actual price range between ~1000 and 4000. Clearly this mean that most of the errors of the regression tree are expected to be quite small, with a few extreme outliers.</p><p>The following plot describes the distribution of the error.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/700/1*n8NuWTlyHUdIcK6-ddOj7Q.png" /></figure><p>Most of the errors have relatively small absolute value (actually near zero in most of the cases) but there are few extreme errors. The MAE is 3.633, where the median absolute error is only 0.83. The really few large absolute errors significantly affect the overall performance. The RMSE in this case is 32.7(!!!), since it gives quadratic strength to the few extremely big errors. Clearly this number say nothing on the typical expected error.</p><p>Although there are some big errors, the model clearly differentiates well between low prices and high prices. The large errors are associated with large actual values of <strong><em>y</em></strong>. Predicting a price of 2,300 instead of actual of 1,800 may be acceptable, where predicting a price of 501, instead of an actual of 1, is unacceptable, yet both cases contribute an absolute error of 500 to the aggregation measure. This implies that measuring the percentage error might be a good idea.</p><p>The MAPE here, however, is 1,129%. Seems to be very bad. The maximal percentage error is over 3,000,000% and it is obtained when y = 3.106549e-05 and the prediction is 1.003 (an absolute error of ~1). The median percentage error is 51%, and the average percentage error when y &gt; 1 is only 32%.</p><p><strong>No single number can tell the entire story</strong></p><p><strong>Communicating performance measures to the business peers</strong></p><p>Where all the above peculiarity might be perfectly clear to data scientists (although we have met so many data scientists failing with these pitfalls), it becomes extremely hard to communicate the complex reality to business peers. No one will accept an average error of 1,000%. No one will accept mean absolute error of 3.5, knowing that most of the sales are closed with prices of less than 1. Failing to communicate the real meaning of a model performance to the business peers is the reason for a no-go decision to POCs that could have been a huge success.</p><p><strong>Practical tips on measuring the performance of regression models</strong></p><p>Here are few practical tips on measuring the performance of regression models.</p><ol><li>Really early in the process (before modeling the data, and even before preparing it for modeling) start a discussion with the business peers on what would make the project a success, business wise. Then try to translate this important business understanding into performance KPIs, and approve these KPIs with the customer. In many cases we find ourselves explaining to the customer the limitations in some suggested KPIs. This way, you don’t have surprises in the validation step, and you can use the right KPIs in the model training step.</li><li>Never use a single static aggregation as the sole performance measure.</li><li>Never use KPIs as black boxes. Remember that error is a numeric random variable, and strive to learn as much as possible on its distribution (is it symmetric? is it correlated with y? does it have extreme values? Is it correlated with any of the explaining features? Etc.). We often use the findings in the analysis of the error that allows us to improve the model (move from linear to non-linear, building an ensemble or a composition of models, and so on).</li><li>Always communicate the meaning of the performance measures to the business users, and help them use this information to tailor a better operational process (e.g., use the model only on specific pre-defined cases).</li></ol><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=817b921a9a62" width="1" height="1" alt=""><hr><p><a href="https://medium.com/yellowblog/evaluating-the-error-of-regression-models-some-real-life-challenges-and-practical-tips-817b921a9a62">Evaluating the Error of Regression Models: Some Real-Life Challenges and Practical Tips</a> was originally published in <a href="https://medium.com/yellowblog">YellowBlog</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The “better than random” model]]></title>
            <link>https://medium.com/yellowblog/the-better-than-random-model-cb503712265e?source=rss-ab3b8d6f6689------2</link>
            <guid isPermaLink="false">https://medium.com/p/cb503712265e</guid>
            <category><![CDATA[machine-learning]]></category>
            <category><![CDATA[heuristics]]></category>
            <category><![CDATA[methodology]]></category>
            <category><![CDATA[risk-management]]></category>
            <dc:creator><![CDATA[Shahar Cohen]]></dc:creator>
            <pubDate>Thu, 13 Jul 2017 07:52:36 GMT</pubDate>
            <atom:updated>2017-07-16T14:10:52.159Z</atom:updated>
            <content:encoded><![CDATA[<p>Although the last think you want to hear about a state of the art machine learning model is that it is just better than random, the way to a winning model often begins with a simple and naive solution, using the most basic modeling techniques.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/960/1*rOKTXS0RjOA3DqbMur3QPA.jpeg" /></figure><p>In machine learning we always get new problems and need to figure out how to solve them. The solution involve research, which is a process that contains significant level of uncertainty. Even the best data science team might fail to deliver a valuable model, simply because valuable is sometimes infeasible. One effective way of hedging the risk is starting with a benchmark model. A good benchmark model have three characteristics:</p><p>i) It is simple and quick to build (quick means faster than a highly handcrafted model, in at least an order of magnitude).</p><p>ii) It has some positive contribution to the task that is being solved (even if it is still not accurate or valuable enough).</p><p>iii) It contains some hints about what is working well and what isn’t working well.</p><p>Due to the 2nd characteristic, we (at <a href="https://medium.com/u/2a6e2e6fd40f">YellowRoad</a>) sometimes call a good benchmark: the “better than random” model.</p><p>Now let’s be clear here. We will always search for a valuable model, business wise, and the “better than random” model is typically just a first baby step on the way. But, in many cases it is a very important baby step. Here are the main reasons for that:</p><ol><li>The “better than random” model is often a good Go/No-Go criterion: If you succeed in building a (simple and quick) “better than random” model, there are high chances that you can invest more, add more parameters (more data, more complexity) and get a significantly better model. It is not a guarantee, but assuming that you attained the “better than random” model easily (significantly faster than the entire project), by having it, you remove a significant amount of risk. We also sometimes see valuable machine learning solutions that were obtained after failing to quickly produce a “better than random” model. But, failing this way significantly increases the risk.</li><li>Running the “better than random” model on real data, characterizing the cases in which it works well, versus the cases in which it fails teaches us a lot about the task that is being solved and helps us on the way to a better model. That is, in many cases, the “better then random” model is not just a risk management step, but a constructive step on the way to a really successful model.</li><li>Ensemble: in machine learning it is sometimes possible to generate an ensemble of relatively weak models, and combine them into a single, surprisingly strong one. If you can generate a series of “better than random” models, you might gain a very good result by combining their sub-results.</li></ol><p>Next time when a customer or a manager asks you how good is the solution, do not answer that it is better than random. None data scientist will probably miss the joke, but let the “better than random” model be part of your working flow.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=cb503712265e" width="1" height="1" alt=""><hr><p><a href="https://medium.com/yellowblog/the-better-than-random-model-cb503712265e">The “better than random” model</a> was originally published in <a href="https://medium.com/yellowblog">YellowBlog</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[What is the most important step in a machine learning project?]]></title>
            <link>https://medium.com/yellowblog/what-is-the-most-important-step-in-a-machine-learning-project-3ee20a13bd20?source=rss-ab3b8d6f6689------2</link>
            <guid isPermaLink="false">https://medium.com/p/3ee20a13bd20</guid>
            <category><![CDATA[machine-learning]]></category>
            <category><![CDATA[methodology]]></category>
            <dc:creator><![CDATA[Shahar Cohen]]></dc:creator>
            <pubDate>Sun, 02 Jul 2017 14:32:38 GMT</pubDate>
            <atom:updated>2017-07-02T14:43:58.329Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/330/1*1JkT0ecHzExbtctMLieBsg.png" /></figure><p><a href="https://en.wikipedia.org/wiki/Cross_Industry_Standard_Process_for_Data_Mining">Cross Industry Standard Process for Data Mining - Wikipedia</a></p><p>The CRISP-DM is a common standard for machine-learning projects. Business Understanding, Data Understanding, Data Preparation, Modeling, Evaluation and Deployment. All these six steps of a machine-learning project are crucial. Quality issues in each step will directly affect the quality of the entire outcome. They are all important.</p><p>However, advising to many organizations on machine learning, and running even more such projects ourselves, we (at YellowRoad) came to a conclusion, that the most under-invested step in the process is Business Understanding. We see many companies discussing algorithms and technology, before understanding the business aspects of the task that they are solving. This is clearly not a good starting point.</p><p>We composed a series of questions that we use in any machine learning project that we get involved in, and we do not invest serious efforts in the following steps, until we have good answers to these questions. We find that practice to be extremely helpful.</p><p>Here are the questions:</p><ol><li>What are we trying to achieve, business wise? Why is it important?</li><li>What are the inputs and outputs for the task that we are trying to solve?</li><li>Given a hypothetical solution to that task, how would it affect our operations? (another way to ask this question: assuming that I have a perfect solution to your machine learning task, how will you use it?)</li><li>Do we already have the ability to act based on such solution, or do we also need to develop that ability? (if the ability is there, learn it carefully. If not, keep close contact with the team that is responsible for developing it)</li><li>How are we going to measure a suggested solution? (KPIs)</li><li>What would make it a success?</li><li>Do we have the input data available? How hard it is to extract it? Are we allowed to use it?</li><li>Are we experience in building similar solutions? Do we understand what it takes?</li><li>Do we have hard budget and timelines constraints?</li><li>Who will develop the solution? Do we have the required skills in house?</li></ol><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=3ee20a13bd20" width="1" height="1" alt=""><hr><p><a href="https://medium.com/yellowblog/what-is-the-most-important-step-in-a-machine-learning-project-3ee20a13bd20">What is the most important step in a machine learning project?</a> was originally published in <a href="https://medium.com/yellowblog">YellowBlog</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
    </channel>
</rss>