Home / Business & Management / Data Science / An Introduction to Artificial Intelligence in the Tourism Industry / Areas of machine learning
Areas of machine learning
You will get an introduction to some machine learning techniques such as supervised learning, unsupervised learning, classification, and clustering.
Machine learning can be divided in various ways. But overall, machine learning is categorized into three areas according to the type of learning. It can be supervised learning ,or unsupervised learning, or reinforcement learning. We are not going to go deep in these areas, but for general information. In supervised learning, the machine learns from label data, which means the data includes the desired output. Supervised learning can be divided into regression ,and classification. In reinforcement learning,
The goal is: we have an agent interacts with its environment by performing actions and learning from errors or when giving him (the agent) a reward, no predefined data is needed and it’s follow trial and error.
In unsupervised learning, We don’t have labeled data. The machine is trained on unlabeled data and training with no supervision. Don’t worry, we will see an example to clarify all these terms. For this video, we will only explain about clustering and classification. To clarify more about labeled and unlabeled data.
For example, we can see in supervised learning the model learned from label data.
When a new unseen data arrived, the model can now use its experience to predict the output of the new data, while in unsupervised learning, we don’t have the desired output. We only have unlabeled data, so instead , the machine learn, it will look at the data and see many variation in that data, then it’ll try to extract features and common attributes, common features, then, cluster or group data according to their similarity. Now we will talk about classification. Classification is a supervised learning approach, which can be thought as a means of categorizing or classifying some unknown items into a set of classes. Classification attempts to learn the relationship between a set of features / between variables and the output variable.
So what is the relation? What are the relationships between our input variables X and our output variable Y?
How the classifier works, you can see here in this table. We have some data, some customer data; feature Variable X and the target variable Y, how the classifier works? first, giving a set of training data point to our model along with the target variable Y, then we train our model to determine the class label for these labeled data point. For example, here we have data and we want to make a model that predict if the customer will have their loan paid or not. First, we gave our model the labeled data from previous experience, for example, previous customers, which include age, gender, income, education for each customer.
Then after training the model in step 2, we can use the trained model for a new customer to predict if we should give this customer /the new customer a loan or not. The popular classification algorithm is the K nearest neighbour. Decision trees and random forest.
Now, we come to clustering. What is clustering? is an unsupervised learning technique, in general, clustering is trying to find group of objects such that the objects in a group will be similar to each other, while the objects in different groups will be different from each other. This is a high level definition. We will not discuss what similar means, just imagine similar means the data point within the group is near from each other, while the data points from different group will be far from each other. Now, let’s take an example. We have this same example, but we have unlabeled data(No target variable). We want to group these data into clusters.
Maybe, for example, cluster 1 will be for rich and middle age customers. Cluster 2 will be for young, educated and middle income customers. Finally, Cluster 3 will be for young and low income customers.
Now the same also, we feed our data to our model, then we can group our data after that when we have un known cluster for a new customer. We simply feed that to our model, then we will have and know which groub this new customer will be belong to. Also, some popular clustering algorithms, such as K-mean algorithm hierarchical clustering and density based clustering.
Machine learning can be categorized into three areas according to the type of learning:
- Supervised
- Unsupervised
- Reinforcement Learning
This video covers these areas.
In supervised learning, the dataset contains the input data (X), which is also called the features, and the output label (Y), and this data is called labeled dataset. We want to train the model on the given dataset (Xi, Yi), then use the trained model to predict a new value of y when giving the model the corresponding x.
- Goal: learn a function that maps input value (x) to output value (y)
- Examples of supervised learning are classification and regression.
The video also introduces.
The video introduces unsupervised learning. In unsupervised learning, we only get raw data (only X) and we don’t have access to the ground truth label (Y). The model is trained with no supervision or guidance.
- Goal: process the data and understands patterns and discovers the outputs
- Examples: clustering, feature extraction
Discussion:
- Discuss and differentiate between supervised and unsupervised learning and lists some of the popular algorithms for classification and clustering.
This article is from the online course:
An Introduction to Artificial Intelligence in the Tourism Industry
Created by
This article is from the free online
An Introduction to Artificial Intelligence in the Tourism Industry
Reach your personal and professional goals
Unlock access to hundreds of expert online courses and degrees from top universities and educators to gain accredited qualifications and professional CV-building certificates.
Join over 18 million learners to launch, switch or build upon your career, all at your own pace, across a wide range of topic areas.
