Machine Learning | Semi-Supervised Learning

Semi-supervised learning is an approach to machine learning that combines a small amount of labeled data with a large amount of unlabeled data during training. Semi-supervised learning falls between unsupervised learning and supervised learning. It is a special instance of weak supervision. #SSL #DataScience #MachineLearning
=====================================================
🐍𝑷𝒚𝒕𝒉𝒐𝒏 𝑺𝒌𝒊𝒍𝒍 𝑺𝒆𝒓𝒊𝒆𝒔 👉 • Skill Series - Python
🐙𝑫𝒆𝒆𝒑 𝑳𝒆𝒂𝒓𝒏𝒊𝒏𝒈 👉 • Deep Learning
🦾𝑴𝒂𝒄𝒉𝒊𝒏𝒆 𝑳𝒆𝒂𝒓𝒏𝒊𝒏𝒈 👉 • Machine Learning
🤖𝑨𝒓𝒕𝒊𝒇𝒊𝒄𝒊𝒂𝒍 𝑰𝒏𝒕𝒆𝒍𝒍𝒊𝒈𝒆𝒏𝒄𝒆 👉 • Artificial Intelligenc...
☁️𝑪𝒍𝒐𝒖𝒅 𝑪𝒐𝒎𝒑𝒖𝒕𝒊𝒏𝒈 👉 • Cloud Computing Tutorials
📶𝑾𝒊𝒓𝒆𝒍𝒆𝒔𝒔 𝑻𝒆𝒄𝒉𝒏𝒐𝒍𝒐𝒈𝒚 👉 • Wireless Technology Tu...
🛠️𝑫𝒂𝒕𝒂 𝑴𝒊𝒏𝒊𝒏𝒈 👉 • Data Mining & Business...
⚙️𝑺𝒊𝒎𝒖𝒍𝒂𝒕𝒊𝒐𝒏 𝑴𝒐𝒅𝒆𝒍𝒊𝒏𝒈 👉 • Simulation Modeling Tu...
🐘𝑩𝒊𝒈 𝑫𝒂𝒕𝒂 👉 • Big Data Anaytics
⛓️𝑩𝒍𝒐𝒄𝒌𝒄𝒉𝒂𝒊𝒏 𝑻𝒆𝒄𝒉𝒏𝒐𝒍𝒐𝒈𝒚 👉 • Blockchain Technology
💡𝑰𝑶𝑻 👉 • Internet Of Things
𝓕𝓸𝓵𝓵𝓸𝔀 𝓶𝓮 𝓸𝓷 𝓘𝓷𝓼𝓽𝓪𝓰𝓻𝓪𝓶 👉 / adhyapakh
𝓥𝓲𝓼𝓲𝓽 𝓶𝔂 𝓟𝓻𝓸𝓯𝓲𝓵𝓮 👉 / reng99
(Feel free to give or ask for any recommendation)
𝓢𝓾𝓹𝓹𝓸𝓻𝓽 𝓶𝔂 𝔀𝓸𝓻𝓴 𝓸𝓷 𝓟𝓪𝓽𝓻𝓮𝓸𝓷 👉 / ranjiraj
𝓖𝓲𝓽𝓗𝓾𝓫👉 github.com/ran...

[@mabasadailycode1781]

As usual, thank you for such clear explanations sir. If possible I would greatly appreciate it if you are in a position to help me understand these kinds of scenarios:
Assuming I have 10000 labeled images then 1000 unlabeled images, SSL would it be an ideal approach or just go with supervised learning since the labeled samples seem to suffice for training a classifier?
Also in a multi-classification problem what if I have a lot of unlabeled data that I believe they are related to some particular classes and not have any possibility of being related to other individual classes from the labeled data. For instance I have 3 classes Dogs, Cats and horses but I have a lot of unlabeled data related to only cats and dogs.
Won’t this imbalance have a negative effect in training my classifier?

[@mabasadailycode1781]

As usual, thank you for such clear explanations sir🙌. If possible I would greatly appreciate it if you are in a position to help me understand these kinds of scenarios:
Assuming I have 10000 labeled images then 1000 unlabeled images, SSL would it be an ideal approach or just go with supervised learning since the labeled samples seem to suffice for training a classifier?
Also in a multi-classification problem what if I have a lot of unlabeled data that I believe they are related to some particular classes and not have any possibility of being related to other individual classes from the labeled data. For instance I have 3 classes Dogs, Cats and horses but I have a lot of unlabeled data related to only cats and dogs.
Won’t this imbalance have a negative effect in training my classifier?

[@peddarapuchandrasekhar8230]

Sir what is label data clarify my dobt

[@mabasadailycode1781]

As usual, thank you for such clear explanations sir🙌. If possible I would greatly appreciate it if you are in a position to help me understand these kinds of scenarios:
Assuming I have 10000 labeled images then 1000 unlabeled images, SSL would it be an ideal approach or just go with supervised learning since the labeled samples seem to suffice for training a classifier?
Also in a multi-classification problem what if I have a lot of unlabeled data that I believe they are related to some particular classes and not have any possibility of being related to other individual classes from the labeled data. For instance I have 3 classes Dogs, Cats and horses but I have a lot of unlabeled data related to only cats and dogs.
Won’t this imbalance have a negative effect in training my classifier?