Skip to content

JA7. Artificial Intelligence and Big Data

Statement

Artificial intelligence has begun to play a big role in data collection and analytics in the big data realm. It provides alternative ways for organizations to capture data and helps reveal patterns not previously seen with other data generation or collection methods.

Identify and describe in detail three ways artificial intelligence can be used with big data analysis.

Answer

Introduction

Artificial intelligence (AI) is a process that involves processing inputs and making decisions in order to increase a performance index; such an index is usually defined before the learning process begins and it tells the model whether it made a good or a bad decision through the designated feedback loop (Benke & Benke, 2018).

In order to make good decisions, AI models must be trained extensively on a huge amount of cases, and this is where big data comes in. Big data provides the framework for AI models to learn and work effectively. There are two AI learning methods: supervised and unsupervised learning. Supervised learning involves training the model against a labeled dataset, while unsupervised learning involves grouping data using clustering algorithms.

The text will discuss three ways AI can be used with big data analysis which are bot detection functionality on a social media app, diagnosis of a disease from pathology results in precision medicine, and the functionality of a recommendation system in an e-commerce platform.

Bot Detection Functionality on a Social Media App

Social media apps are platforms where millions of human users interact. However, the number of bots increased recently which refers to social media accounts that do not represent a human user, but rather an automated script that is used to perform certain tasks, good or bad. Bots were recently used for malicious purposes such as spreading fake news, spamming, and manipulating public opinion.

AI can be used to make decisions about whether an account is a bot or a human. AI can help in this process by being trained on a large dataset of human interactions with the app. For example, a machine learning algorithm such as a neural network could be trained to detect certain recurring words, phrases, and views expressed in social media and then use pattern recognition to correlate the data with news content, candidate names, media outlets, and geographic locations. Suspicious correlations and clusters can be revealed in this manner (Benke & Benke, 2018).

The performance index of such a model would be how close an account’s behavior is to the average human behavior that the model has detected during training, and this is a way of unsupervised learning. The risk of flagging a human account as a bot is always present, but the model can be improved by human intervention which reverses the decision and provides feedback to the model.

Precision Medicine: Diagnosis of a Disease from Pathology Results

In precision medicine, the goal is to provide the right treatment at maximum accuracy; diagnosing the right disease is the first step in the process. There are many factors affecting the diagnosis process such as the patient’s genetic information, lifestyle, other diseases, and the information available in the medical literature.

AI can help in analyzing billions of medical records and pathology results to flag certain patterns for human experts to look into in an unsupervised learning manner. It can also make decisions if a certain disease exists or not based on a supervised learning model where experts feed the model with positive and negative cases of the disease, and the model performs detection, discrimination, and classification tasks to make a decision.

Recommendation System in an E-commerce Platform

Recommendation systems aim to create more revenue by providing the user with a personalized list of products that they might be interested in buying. AI can be trained on a large dataset of the history of transactions and cross it with user information to learn rules and patterns that can be used to recommend products to users in an unsupervised learning manner.

After training, the AI model keeps track of the recommendations it produces and monitors users’ reactions to them; this should create a feedback loop that helps the model improve its performance index in a supervised learning manner. The performance index in this case is how many users bought the recommended product, and how many users did not buy it.

Conclusion

AI and big data are two technologies that are intertwined and complement each other. To make AI models useful, they must be trained on big data sets. AI can be used in many ways with big data analysis using supervised and unsupervised learning methods. AI can help in distinguishing between human and bot accounts on social media apps, diagnosing diseases from pathology results in precision medicine, and recommending products to users in an e-commerce platform.

References