{"id":1851,"date":"2020-09-23T14:50:01","date_gmt":"2020-09-23T14:50:01","guid":{"rendered":"https:\/\/data-science.gotoauthority.com\/2020\/09\/23\/how-i-consistently-improve-my-machine-learning-models-from-80-to-over-90-accuracy\/"},"modified":"2020-09-23T14:50:01","modified_gmt":"2020-09-23T14:50:01","slug":"how-i-consistently-improve-my-machine-learning-models-from-80-to-over-90-accuracy","status":"publish","type":"post","link":"https:\/\/wealthrevelation.com\/data-science\/2020\/09\/23\/how-i-consistently-improve-my-machine-learning-models-from-80-to-over-90-accuracy\/","title":{"rendered":"How I Consistently Improve My Machine Learning Models From 80% to Over 90% Accuracy"},"content":{"rendered":"
\n

By Terence Shin<\/a>, Data Scientist | MSc Analytics & MBA student<\/b>.<\/p>\n

<\/p>\n

Photo by\u00a0Ricardo Arce<\/a>\u00a0on\u00a0Unsplash<\/a>.<\/em><\/p>\n

If you\u2019ve completed a few data science projects of your own, then you probably realized by now that achieving an accuracy of 80% isn\u2019t too bad! But in the real world, 80% won\u2019t cut it. In fact, most companies that I\u2019ve worked for expect a minimum accuracy (or whatever metric they\u2019re looking at) of at least 90%.<\/p>\n

Therefore, I\u2019m going to talk about 5 things that you can do to significantly improve your accuracy.\u00a0I highly recommend that you read all five points thoroughly<\/strong>\u00a0because there are a lot of details that I\u2019ve included that most beginners don\u2019t know.<\/p>\n

By the end of this, you should understand that there are many more variables than you think that play a role in dictating how well your machine learning model performs.<\/p>\n

With that said, here are 5 things that you can do to improve your machine learning models!<\/p>\n

\u00a0<\/p>\n

1. Handling Missing Values<\/h3>\n

\u00a0<\/p>\n

One of the biggest mistakes I see is how people handle missing values, and it\u2019s not necessarily their fault. A lot of material on the web says that you typically handle missing values through\u00a0mean imputation,\u00a0<\/strong>replacing null values with the mean of the given feature, and this usually isn\u2019t the best method.<\/p>\n

For example, imagine we have a table showing age and fitness score and imagine that an eighty-year-old has a missing fitness score. If we took the average fitness score from an age range of 15 to 80, then the eighty-year-old will appear to have a much higher fitness score that they actually should.<\/p>\n

Therefore, the first question you want to ask yourself is\u00a0why<\/strong>\u00a0the data is missing to begin with.<\/p>\n

Next, consider other methods in handling missing data, aside from mean\/median imputation:<\/p>\n