What is feature engineering?
Feature engineering is the process of using domain knowledge to extract features (characteristics, properties, attributes) from raw data. A feature is a property shared by independent units on which analysis or prediction is to be done. Features are used by predictive models and influence results.
What is the purpose of feature engineering?
When your goal is to get the best possible results from a predictive model, you need to get the most from what you have. This includes getting the best results from the algorithms you are using. It also involves getting the most out of the data for your algorithms to work with.
This is the problem that the process and practice of feature engineering solves.
The features in your data will directly influence the predictive models you use and the results you can achieve.
You can say that: the better the features that you prepare and choose, the better the results you will achieve. It is true, but it also misleading.
The results you achieve are a factor of the model you choose, the data you have available and the features you prepared. Even your framing of the problem and objective measures you’re using to estimate accuracy play a part. Your results are dependent on many inter-dependent properties. You need great features that describe the structures inherent in your data.
Is feature engineering good?
Better features means flexibility
You can choose “the wrong models” (less than optimal) and still get good results. Most models can pick up on good structure in data. The flexibility of good features will allow you to use less complex models that are faster to run, easier to understand and easier to maintain. This is very desirable.
Better features means simpler models.
With well engineered features, you can choose “the wrong parameters” (less than optimal) and still get good results, for much the same reasons. You do not need to work as hard to pick the right models and the most optimized parameters.
With good features, you are closer to the underlying problem and a representation of all the data you have available and could use to best characterize that underlying problem.
What does feature engineering include?
The process involves a combination of data analysis, applying rules of thumb, and judgement. It is sometimes referred to as pre-processing, although that term can have a more general meaning.
What is feature engineering example?
The most common type of data is continuous data. It can take any values from a given range. For example, it can be the price of some product, the temperature in some industrial process or coordinates of some object on the map.
Feature generation here relays mostly on the domain data. For example, we can subtract the warehouse price from the shelf one to calculate the profit or we can calculate the distance between two locations on the map.
The new possible features are limited only by the available features and known mathematical operations.
What are feature engineering techniques?
Missing values are one of the most common problems you can encounter when you try to prepare your data for machine learning. The reason for the missing values might be human errors, interruptions in the data flow, privacy concerns, and so on. Whatever is the reason, missing values affect the performance of the machine learning models.
The most simple solution to the missing values is to drop the rows or the entire column. There is not an optimum threshold for dropping but you can use 70% as an example value and try to drop the rows and columns which have missing values with higher than this threshold.
2. Handling Outliers
The best way to detect the outliers is to demonstrate the data visually. All other statistical methodologies are open to making mistakes, whereas visualizing the outliers gives a chance to take a decision with high precision.
Statistical methodologies are less precise, but on the other hand, they have a superiority, they are fast. There are a few different ways to handle outliers:
- Standard deviation
- Drop or cap
The main motivation of binning is to make the model more robust and prevent overfitting, however, it has a cost to the performance. Every time you bin something, you sacrifice information and make your data more regularized. Binning can be applied on both categorical and numerical data.
4. Log Transform
Logarithm transformation (or log transform) is one of the most commonly used mathematical transformations in feature engineering. What are the benefits of log transform:
- It helps to handle skewed data and after transformation, the distribution becomes more approximate to normal.
- In most of the cases the magnitude order of the data changes within the range of the data. For instance, the difference between ages 15 and 20 is not equal to the ages 65 and 70. In terms of years, yes, they are identical, but for all other aspects, 5 years of difference in young ages mean a higher magnitude difference. This type of data comes from a multiplicative process and log transform normalizes the magnitude differences like that.
- It also decreases the effect of the outliers, due to the normalization of magnitude differences and the model become more robust.
5. One-Hot Encoding
One-hot encoding is one of the most common encoding methods in machine learning. This method spreads the values in a column to multiple flag columns and assigns 0 or 1 to them. These binary values express the relationship between grouped and encoded column.
This method changes your categorical data, which is challenging to understand for algorithms, to a numerical format and enables you to group your categorical data without losing any information.
6. Grouping Operations
In most machine learning algorithms, every instance is represented by a row in the training dataset, where every column show a different feature of the instance. This kind of data called “Tidy.”
Datasets such as transactions rarely fit the definition of tidy data above, because of the multiple rows of an instance. In such a case, we group the data by the instances and then every instance is represented by only one row.
7. Feature Split
Splitting features is a good way to make them useful in terms of machine learning. Most of the time the dataset contains string columns that violates tidy data principles. By extracting the utilizable parts of a column into new features:
We enable machine learning algorithms to comprehend them.
- Make possible to bin and group them.
- Improve model performance by uncovering potential information.
Split function is a good option, however, there is no one way of splitting features. It depends on the characteristics of the column, how to split it.
In most cases, the numerical features of the dataset do not have a certain range and they differ from each other. In real life, it is nonsense to expect age and income columns to have the same range. But from the machine learning point of view, how these two columns can be compared?
Scaling solves this problem. The continuous features become identical in terms of the range, after a scaling process. This process is not mandatory for many algorithms, but it might be still useful to apply.
9. Extracting Date
Though date columns usually provide valuable information about the model target, they are neglected as an input or used nonsensically for the machine learning algorithms. It might be the reason for this, that dates can be present in numerous formats, which make it hard to understand by algorithms, even they are simplified to a format like "01–01–2017".
Building an ordinal relationship between the values is very challenging for a machine learning algorithm if you leave the date columns without manipulation. There are three types of preprocessing for dates:
- Extracting the parts of the date into different columns: Year, month, day, etc.
- Extracting the time period between the current date and columns in terms of years, months, days, etc.
- Extracting some specific features from the date: Name of the weekday, Weekend or not, holiday or not, etc.
If you transform the date column into the extracted columns like above, the information of them become disclosed and machine learning algorithms can easily understand them.