1) What is Machine learning?
Machine learning is the field of study that gives computers the ability to learn without being explicitly programmed - Arthur Samuel,1959.
2)What is Data Mining?
Digging large amounts of data to discover hidden patterns is called data mining
3) Types of Machine Learning
Supervised Learning: Here, the model is trained on labeled data. An example of labeled data is data that has specifications about each item, such as "This is a dog image."
Unsupervised Learning: Here, the model is trained on unlabeled data.
Semi-supervised Learning: First, the model is trained on a small amount of labeled data. Based on that, it is then trained on a large amount of unlabeled data. This comes into the picture when labeled data is expensive to obtain.
Self-supervised Learning: The model generates its own labels for the data.
Reinforcement Learning: An agent learns by interacting with the environment through trial and error.
4)What is Visualization Algorithms?
Visualization algorithms are a good example of Unsupervised Learning. Here, the model is given a large amount of unlabeled and complex data, and the output is a 2D or 3D representation that helps us understand patterns, relationships, and clusters in the data
5)What dimensionality reduction ?
Dimensionality Reduction is a technique used to simplify data by reducing the number of features while keeping most of the important information. It combines or removes related features to make the data smaller and easier to understand, with very little loss of information.
6)Association Rule learning?
Association Rule Learning is an unsupervised learning technique used digging into large data sets, to identify relation between attributes.
in a supermarket, the algorithm may find that customers who purchase milk often purchase bread as well. This helps businesses understand customer buying patterns.
7)What is instance based Learning?
Instance based learning memorize the training data, based on trained data example will give response to inputs.
8)What is Model based learning?
Model based learning creates a generalized model from the trained data uses that model for predictions
9)How would you train a robot to walk on unfamiliar terrain using machine learning?
Reinforcement Learning
10)What is online learning system?
This is the one of the AI/ML approach where the model is trained continuously when new data arrives
Example : supermarket stock details, Netflix and YouTube video recommendations
11)What is out of core learning?
Training a machine learning model on a dataset that is too large to fit into a single machine's main memory (RAM). Instead of loading the entire dataset, the algorithm loads data in small chunks incrementally and updates the model accordingly.
It is similar to online learning; the difference is that online learning algorithms learn incrementally as new data arrives, whereas out-of-core learning processes an existing dataset batch by batch.
12)What are the major challenges faced while working with machine learning?
Insufficient data
Non representative data
poor quality
irrelevant feature
overfitting
underfitting
13)If your model performs very well on the training data but generalizes poorly to new instances, what could be the reason?
The model has learned the training data very well rather than learning true underlying pattern. So its performs very well on data it has already seen, but fails to generalize to new, unseen instances.
Common causes:
Complex model.
little training data.
Training data has errors/Noisy
Too many features
Solutions:
Simplify the model Gather more training data Reduce noise in the data Reduce number of features Apply regularization Increase regularization hyperparameter
14)What is Training Set, Validation Set (Dev Set), Test Set and Train-Dev Set ?
Training Set: This data is used to train the model
Validation Set (Dev Set): Tune hyperparameters, model selection , The model never trains on this data directly
Test Set: Used only once, at the very end, to evaluate the final model's performance on completely unseen data to understand how model is working real time prod data.
Train-Dev Set: A subset taken from / split off from the training data (same distribution as training), used to distinguish overfitting from data mismatch
15)What is Overfitting and Underfitting in AI/ML?
Overfitting : Model trained very well on training data (including noisy) , so gives good performance on training data , poor performance on new data.
Underfitting: Model is too simple to capture underlying pattern so poor performance on training and new data.
16)What is Median? or మధ్యగతం అంటే ఏమిటి?
Median is the middle value in a data set or series of numbers when the numbers are arranged in order (either ascending or descending). If there's an even number of values, the median is the average of the two middle numbers.
Telugu: మధ్యగతం (madhyagatam) or మధ్యస్థం (madhyasthaṁ)
మధ్యగతం (madhyagatam) is the more commonly used statistical term for "median."
It literally relates to "మధ్య" (madhya) meaning "middle."
2, 5, 7, 9, 15
ఇక్కడ మొత్తం 5 సంఖ్యలు ఉన్నాయి.
మధ్యలో ఉన్న సంఖ్య 7.
17)రిగ్రెషన్ (Regression) అంటే ఏమిటి? ఎన్ని రకాలు (or) What is Regression and Types of Regression?
రిగ్రెషన్ను రెండు రకాలుగా చెప్పవచ్చు — ఒకటి మెథడ్ ఆధారంగా (Linear, Logistic, etc.), రెండోది అవుట్పుట్ వేరియబుల్స్ సంఖ్య ఆధారంగా (Univariate vs Multivariate).
ఈ రెండు ఇండిపెండెంట్ డైమెన్షన్లు.
Regression can be classified as two — one based on the modeling method (Linear, Logistic, etc.), and another based on the number of output variables (Univariate vs Multivariate).
These are two independent dimensions of classification.
No comments:
Post a Comment