Yes, machine learning or Artificial Intelligence sounds like a futuristic sci-fi movie, where a robot serves you breakfast and drives your car. Many believe it’s a far-fetched experiment from scientists at Silicon Valley.
The truth is, you use machine learning all the time, but you probably don’t know where.
Let’s take the example of this website.
You may have reached this blog by a Google search or a social media post, but the search is a result of your machine learning about your preferences, behaviors and networks of friends and acquaintances. The machine or algorithm has learned from you, to provide you with valuable information, and will continue to improve this.
What is Machine Learning?
According to Stanford University, Machine Learning is a field that concentrates on induction and other types of algorithms that can be used to “learn.”
What does this mean? You input information on a daily basis (whether you’re aware of it or not), and the software learns from the data.
Machine Learning can be applied to:
- Fraud detection.
- Web search results.
- Real-time ads on web pages and mobile devices.
- Text-based sentiment analysis.
- Credit scoring and next-best offers.
- Prediction of equipment failures.
- New pricing models.
- Network intrusion detection.
- Pattern and image recognition.
- Email spam filtering
and much more…
What’s the difference between data mining and machine learning?
According to SAS, “while machine learning uses many of the same algorithms and techniques as data mining, the difference lies in what issues the two disciplines predict.”
- Data mining discovers previously unknown patterns and knowledge.
- Machine learning is used to reproduce known patterns and knowledge, automatically apply that to other data, and then automatically apply those results to decision making and actions.
Machine learning and higher education retention strategies
Students who drop out from college and fail to graduate are an economical and emotional loss for families and institutions. That is why higher education institutions are using methods and customized support from data, studies, forms and specialized literature.
What if you were able to alert student defection by machine learning and project your demand of courses and groups, planning and controlling assistance?
According to Dursun Delen, from the Oklahoma State University, “Student retention is an essential part of many enrollment management systems. It affects university rankings, school reputation and financial wellbeing.”
For educational software experts like him, “Improving student retention starts with a thorough understanding of the reasons behind the attrition. Such an understanding is the basis for accurately predicting at-risk students and appropriately intervening to retain them.”
Professor Delen studied five years of institutional data and developed analytical models to predict and to explain the reasons behind freshmen student attrition. He found that educational and financial variables, in his sample, were the most significant predictors for attritions.
This has developed into practical implementations.
With the right algorithm and the proper input, machine learning can help:
- Detect on an early stage student with high levels of attrition.
- Identify most recurrent risk factors.
- Report the main causes and risk factors.
How can a machine learn from my students?
From the moment a student expresses interest in joining an undergraduate education, they are providing insightful information about their interests, pain points and priorities.
At the time they enroll at a given institution, students and parents input several other data, such as house income, address, class schedule and grades.
During their time at school, students, teachers and administrators input extraordinary amount of information either directly –through forms or emails – and indirectly, through the use of external databases.
A practical case
Imagine the case of Pedro.
- He got accepted in Mechanical Engineering at one of the best universities of the world, located at a massive campus in central Mexico, with a population of over 40,000 students.
- On his application, he stated that Mechanics was his second choice, after he was rejected from Computer Engineering at the same institution.
- He began getting excellent grades and won a half tuition scholarship. Furthermore, his library records show he spent substantial amount of time at the library browsing through selected books.
However, from the second year of school, something happened:
- His time and attendance records dropped: he began skipping lectures on the eve of weekends.
- His records showed that he began living at a student residence, as his hometown was over 700km away from campus.
- His financial form provided evidence that his parents had used a loan to pay for education, but that over previous months these payments were delayed. Pedro may be at risk of dropping out due to financial and geographical reasons.
From over 5,000 records of students at different levels of risk, before beginning year three, Student Affairs got an alert on Pedro’s situation, as he hadn’t yet decided on his credits and had skipped the pre-sessional courses.
A student advisor called him to the office.
It was true, Pedro’s family was going through a financial crisis, and he was forced to work on weekends – including Fridays –only to pay for his student residence. He was, in fact, considering dropping out of college.
Later, the university presented Pedro with an opportunity. He wouldn’t be able to attend Mechanical Engineering in the city, but – a couple of years before – the University had opened a small campus only an hour away from his hometown. One of the majors was Computer Engineering.
Even though the university had rejected him in the first round, with his academic track record, he was advised to internally switch majors, as he had proven worthy of an early bird registration on the local campus program.
The advisor had no clue about Pedro’s track record. It was confidential. He only got selective information after the system alerted that he was at serious risk of dropping out.
How did the alert know about this? Because it was configured to detect these patterns, as academic literature had evidence that students with such dramatic change in their academic records were at risk of leaving school.
These solutions and compromises tend to happen in many universities. However, most higher education institutions lack the input to react responsively to situations that may prevent losing talent in universities. What if the institutions could do something about it?
Do you think machine learning is a good tool for higher education? Do you have concerns about the use of these applications?