Categories: Salaries in Analytics

# Decision Tree in Machine Learning: Types, Advantages, Disadvantages in 5 Points

## Introduction

When you have an issue with your internet connection and dial-up customer care, an Intelligent computer assistant is the one you are first connected to. After dialling a bunch of numbers to make decisions as to what you seek, you are finally connected to a human support system. One may often think this to be just standardized voicemail; in reality, it is a real-life example of what a decision tree in machine learning entails; helping reach the right choice.

## 1) Definition

Now, what is a decision tree? Decision Tree is a useful machine learning program that can be used for solving both classification and regression problems. They are powerful analytical models that have the ability to comprehend data with minimal pre-processing time. It is a support tool that has a tree-like structure and suggests possible effects and costs of decisions.

## 2) Construction

A decision tree presents an algorithm regarding decision making in a flowchart like structure.

A tree can be understood by dividing the source set into subsets depending on a quality worth test. This cycle is rehashed on each determined subset in a recursive way known as recursive partitioning. The recursion is said to be complete when the subset at the node is the same value as that of the target variable subset or when splitting does not add any more value to the predictions made. The development/ construction of a decision tree classifier does not need any knowledge of the domain or boundary setting.

Decision trees can deal with high dimensional data. All in all, decision trees classifier has great precision and an inductive approach to imbibe knowledge on characterization.

There are often assumptions that are made while creating a decision tree. Some of them are:

• In the initial stages, the whole set is considered as the root.
• The values need to be categorical and continuous till it is used to build the model.
• The records have to be distributed recursively.
• A statistical approach needs to be adopted while placing attributes as root or internal node of the tree.

## 3) Representation

Decision trees group or classify occurrences by arranging them down the tree; from root to different leaves nodes. This provides the occurrences’’ characterization or classification. An occurrence is classified by beginning at the root node of the tree, testing the traits indicated by this node, then progressing down the tree branch by comparing it to the attribute’s value. The process is repeated for the subtree that is rooted at the new node.

## 4) Examples

The following is an example of a binary tree model. Suppose you want to anticipate whether an individual is fit given that they have provided essential data such as their age, dietary pattern, physical activities, etc. The decision nodes here would be questions like, “what’s the age”, “does he work out?”, “Does he eat too many pizzas?”. What’s more, the leaves, which are essentially the results, are either “fit” or “unfit”. The two answers make it binary classification.

Another example of binary classification would be deciding on whether circumstances are right to play tennis one morning. The instances of outlook ( that is, rain or sunny), temperature (hot or cold), humidity (high or low), wind (strong or not), etc. will form branches of the decision tree. A disjunction of different collected constraints will provide attributed value to the instance in question and in the present case, the instance being suitable weather to play tennis.

In comparison to various decision-making tools, decision trees have several advantages. Some of the are:

• While utilizing a decision tree algorithm, it is not essential to standardize or normalize the data that has been collected. It can handle both continuous and categorical variables.
• The execution of a Decision tree algorithm must be possible without having to scale the data as well.
• While utilizing the decision tree algorithm, it is not necessary to credit the missing values.
• Unlike the traditional pre-processing steps of data, the pre-processing steps require lesser coding and analysis in a decision tree making model.
• Unlike the traditional pre-processing steps of data, the pre-processing steps are time-saving in a decision tree making model.
• Comprehensive rules are generated in a decision tree.
• The idea/ concept that drives the decision tree making model is more familiar and easier for developers/ programmers in comparison to other algorithms.

However, as there are pros, there are cons to the decision tree making models as well. Some of them are:

• The numerical calculations involved in a decision tree generally consumes more memory.
• The numerical estimations made in a decision tree requires additional time.
• The reproducibility of a decision tree is exceptionally sensitive as even a slight change in data can bring about an enormous change in the tree structure.
• The multifaceted space and time complexity in a decision tree model are generally higher than other decision-making tools.
• Due to this multifaceted complexity, the training time involved in Decision tree model is higher as well. This increases the cost involved in training as well.
• A single decision tree model is a frail learner, thus creating a requirement for multiple decision trees which can then provide better predictions.
• The drawbacks of decision tree making models make it less appropriate for tasks that need to predict continuous attribute values.
• It is quite expensive to create a decision tree as, at each node, the sorting of fields must be done. In certain algorithms, a combination of different fields is used simultaneously, increasing costs even more. Pruning algorithms too is not easy on the pocket because many candidate subtrees may be formed and compared with each other.

## Conclusion

From the above discussion, it is clear that decision trees can handle non-linear data sets in an effective manner. It plays as a decision-making catalyst in various life walks, including engineering, civil planning, business, and even law. On perusal of the advantages and disadvantages involved in adopting a decision tree model must be done based on its suitability to the problem statement at hand.

The basic details of Deep Neural Network are covered in this article. The article is a guide on Deep Neural Network learning for beginners. Upskilling to emerging technologies has become the need of the hour, with technological changes shaping the career landscape. If you are interested in learning more about Artificial Intelligence & Deep Learning, Jigsaw Academy’s 6-month online Postgraduate Certificate Program in AI & Deep Learning will help you upskill, stay relevant, and get noticed.

## Books on Analytics

Analytics is a vast field. At the one end, it overlaps with statistics and higher…

## Career in analytics in a KPO

Do you love to explore and investigate information? Do you find spreadsheets to be a…

## Indian companies using analytics

India has developed into the global hub for analytics. A large number of MNCs have…

## IBM: Betting big on analytics

International Business Machines Corp. Or IBM as it is popularly known recently announced its restructuring…

## How to build a successful career in analytics

So you have got a job as an analyst in your dream company? Here are…

## What’s the sentiment on “sentiment analysis”?

What's the sentiment on "sentiment analysis"? Is the field ready to take off?