Forum Discussion
How to explain on what AI / ML can do?
As a data analyst, when client expects AI / ML, since general people don't understand much what AI / ML is about, so what do they usually expect and explain to them?
Does anyone have any suggestions?
Thanks in advance
oem7110 , You can consume the Azure model
https://learn.microsoft.com/en-us/power-bi/connect-data/service-aml-integrate
Python- https://www.datacamp.com/tutorial/running-python-scripts-in-power-bi-tutorial
Usually, both Machine learning and deep learning are Supervised - which means the data we train has the result/outcome with it. And Unsupervised - This means there is no reference result/outcome.
Then we have a mix - Semi-Supervised. And reinforcement learning when the algorithm learns with each new input.
Linear regression, Logistic regression, etc are supervised ML. Because you have the end results with data.
https://www.ibm.com/in-en/cloud/learn/machine-learning
4 Replies
- amitchandak
Super User
oem7110 , refer if this can help
https://www.pcmag.com/news/the-business-guide-to-machine-learning
"Deep learning is where much of the innovation around neural networks lies."
Linear regression is one of machine learning algorithm based on supervised learning, which is also available under Excel.
If there are 10 monitoring items, traditionally, I can create a 2D matrix to find correlation for any 2 pairs, that is not user friendly, Under Power BI, is there any tools avaiable to input 10 items? then it would generate report on their correlation.
Is there any video sample to introduce on how Power BI implements ML?
Do you have any suggestions?
Thank you very much for any suggestions (^v^)- amitchandak
oem7110 , You can consume the Azure model
https://learn.microsoft.com/en-us/power-bi/connect-data/service-aml-integrate
Python- https://www.datacamp.com/tutorial/running-python-scripts-in-power-bi-tutorial
Usually, both Machine learning and deep learning are Supervised - which means the data we train has the result/outcome with it. And Unsupervised - This means there is no reference result/outcome.
Then we have a mix - Semi-Supervised. And reinforcement learning when the algorithm learns with each new input.
Linear regression, Logistic regression, etc are supervised ML. Because you have the end results with data.
https://www.ibm.com/in-en/cloud/learn/machine-learning
"There's a large misconception that you can just throw your data into a product and get some beautiful insights out of it."
How to explain clients about this misconception?
Do you have any suggestions?
Thank you very much for any suggestions (^v^)
