Machine Learning Cheatsheet

#python #ml #cheatsheet #ai

Pattern Recognition and
Machine Learning [ Information Science and Statistics ]

Christopher M. Bishop
#python #machinelearning #statistics #information #ai #ml

📕 Introduction to Machine Learning
by Alex Smola and S.V.N. Vishwanathan

University Press, Cambridge

#numpy

NumPy

Smart use of ‘:’ to extract the right shape


Sometimes you encounter a 3-dim array that is of shape (N, T, D), while your function requires a shape of (N, D). At a time like this, reshape() will do more harm than good, so you are left with one simple solution:

Example:

for t in xrange(T):
x[:, t, :] = # ...

To become a Machine Learning Engineer:

• Python
• numpy, pandas, matplotlib, Scikit-Learn
• TensorFlow or PyTorch
• Jupyter, Colab
• Analysis > Code
• 99%: Foundational algorithms
• 1%: Other algorithms
• Solve problems ← This is key
• Teaching = 2 × Learning
• Have fun!

A LITTLE GUIDE TO HANDLING MISSING DATA
Having any Feature missing more than 5-10% of its values? you should consider it to be missing data or feature with high absence rate👀

How can you handle these missing values, ensuring you dont loose important part of your data🤷‍♀️
Not a problem😌. Here are important facts you must know😉

✍️Instances with missing values for all features should be eliminated
✍️Features with high absence rate should either be eliminated or filled with values
✍️Missing values can be replaced using Mean Imputation or Regression Imputation
✍️ Be careful with mean imputation for it may introduce bias as it evens out all instances
✍️Regression Imputation might overfit your model
✍️Mean and Regression Imputation can't be applied to Text features with missing values
✍️Text Features with missing values can be eliminated if not needed in data
✍️Important Text Features with Missing values can be replaced with a new class or category labelled as uncategorized

Top 8 Github Repos to Learn Data Science and Python

1. All algorithms implemented in Python
By: The Algorithms
Stars ⭐️: 135K
Fork: 35.3K
Repo: https://github.com/TheAlgorithms/Python

2. DataScienceResources
By: jJonathan Bower
Stars ⭐️: 3K
Fork: 1.3K
Repo: https://github.com/jonathan-bower/DataScienceResources

3. Playground and Cheatsheet for Learning Python
By: Oleksii Trekhleb ( Also the Image)
Stars ⭐️: 12.5K
Fork: 2K
Repo: https://github.com/trekhleb/learn-python


4. Learn Python 3
By: Jerry Pussinen
Stars ⭐️: 4,8K
Fork: 1,4K
Repo: https://github.com/jerry-git/learn-python3


5. Awesome Data Science
By: Fatih Aktürk, Hüseyin Mert & Osman Ungur, Recep Erol.
Stars ⭐️: 18.4K
Fork: 5K
Repo: https://github.com/academic/awesome-datascience

6. data-scientist-roadmap
By: MrMimic
Stars ⭐️: 5K
Fork: 1.5K
Repo: https://github.com/MrMimic/data-scientist-roadmap


7. Data Science Best Resources
By: Tirthajyoti Sarkar
Stars ⭐️: 1.8K
Fork: 717
Repo: https://github.com/tirthajyoti/Data-science-best-resources/blob/master/README.md


8. Ds-cheatsheets
By: Favio André Vázquez
Stars ⭐️: 10.4K
Fork: 3.1K
Repo: https://github.com/FavioVazquez/ds-cheatsheets

💥Deep Learning with Pytorch by Prof.Yann LeCun (CNN Founder)

This course concerns the latest techniques in deep learning and representation learning, focusing on supervised and unsupervised deep learning, embedding methods, metric learning, convolutional and recurrent nets, with applications to computer vision, natural language understanding, and speech recognition.

GitHub Link: https://atcold.github.io/pytorch-Deep-Learning/

YouTube Playlist: https://www.youtube.com/playlist?list=PLLHTzKZzVU9eaEyErdV26ikyolxOsz6mq

Probability Cheat Sheet
👇👇
https://web.cs.elte.hu/~mesti/valszam/kepletek

New Data Scientists - When you learn, it's easy to get distracted by Machine Learning & Deep Learning terms like "XGBoost", "Neural Networks", "RNN", "LSTM" or Advanced Technologies like "Spark", "Julia", "Scala", "Go", etc.

Don't get bogged down trying to learn every new term & technology you come across.

Instead, focus on foundations.
- data wrangling
- visualizing
- exploring
- modeling
- understanding the results.

The best tools are often basic, Build yourself up. You'll advance much faster. Keep learning!

Data Analysis Interview Questions and Answers
👇👇

1.How to create filters in Power BI?

Filters are an integral part of Power BI reports. They are used to slice and dice the data as per the dimensions we want. Filters are created in a couple of ways.

Using Slicers: A slicer is a visual under Visualization Pane. This can be added to the design view to filter our reports. When a slicer is added to the design view, it requires a field to be added to it. For example- Slicer can be added for Country fields. Then the data can be filtered based on countries.
Using Filter Pane: The Power BI team has added a filter pane to the reports, which is a single space where we can add different fields as filters. And these fields can be added depending on whether you want to filter only one visual(Visual level filter), or all the visuals in the report page(Page level filters), or applicable to all the pages of the report(report level filters)


2.How to sort data in Power BI?

Sorting is available in multiple formats. In the data view, a common sorting option of alphabetical order is there. Apart from that, we have the option of Sort by column, where one can sort a column based on another column. The sorting option is available in visuals as well. Sort by ascending and descending option by the fields and measure present in the visual is also available.


3.How to convert pdf to excel?

Open the PDF document you want to convert in XLSX format in Acrobat DC.
Go to the right pane and click on the “Export PDF” option.
Choose spreadsheet as the Export format.
Select “Microsoft Excel Workbook.”
Now click “Export.”
Download the converted file or share it.


4. How to enable macros in excel?

Click the file tab and then click “Options.”
A dialog box will appear. In the “Excel Options” dialog box, click on the “Trust Center” and then “Trust Center Settings.”
Go to the “Macro Settings” and select “enable all macros.”
Click OK to apply the macro settings.

————————————————————-

ENJOY LEARNING 👍👍

Some helpful Data science projects for beginners

https://www.kaggle.com/c/house-prices-advanced-regression-techniques

https://www.kaggle.com/c/digit-recognizer

https://www.kaggle.com/c/titanic

Intermediate Level Data science Projects

Black Friday Data : https://www.kaggle.com/sdolezel/black-friday

Human Activity Recognition Data : https://www.kaggle.com/uciml/human-activity-recognition-with-smartphones

Trip History Data : https://www.kaggle.com/pronto/cycle-share-dataset

Million Song Data : https://www.kaggle.com/c/msdchallenge

Census Income Data : https://www.kaggle.com/c/census-income/data

Movie Lens Data : https://www.kaggle.com/grouplens/movielens-20m-dataset

Twitter Classification Data : https://www.kaggle.com/c/twitter-sentiment-analysis2

Text mining : https://www.kaggle.com/kanncaa1/applying-text-mining

Three different learning styles in machine learning algorithms:

1. Supervised Learning

Input data is called training data and has a known label or result such as spam/not-spam or a stock price at a time.

A model is prepared through a training process in which it is required to make predictions and is corrected when those predictions are wrong. The training process continues until the model achieves a desired level of accuracy on the training data.

Example problems are classification and regression.

Example algorithms include: Logistic Regression and the Back Propagation Neural Network.

2. Unsupervised Learning

Input data is not labeled and does not have a known result.

A model is prepared by deducing structures present in the input data. This may be to extract general rules. It may be through a mathematical process to systematically reduce redundancy, or it may be to organize data by similarity.

Example problems are clustering, dimensionality reduction and association rule learning.

Example algorithms include: the Apriori algorithm and K-Means.

3. Semi-Supervised Learning

Input data is a mixture of labeled and unlabelled examples.

There is a desired prediction problem but the model must learn the structures to organize the data as well as make predictions.

Example problems are classification and regression.

Example algorithms are extensions to other flexible methods that make assumptions about how to model the unlabeled data.

Interview QnAs For ML Engineer

1.What are the various steps involved in an data analytics project?

The steps involved in a data analytics project are:

Data collection
Data cleansing
Data pre-processing
EDA
Creation of train test and validation sets
Model creation
Hyperparameter tuning
Model deployment


2. Explain Star Schema.

Star schema is a data warehousing concept in which all schema is connected to a central schema.


3. What is root cause analysis?

Root cause analysis is the process of tracing back of occurrence of an event and the factors which lead to it. It’s generally done when a software malfunctions. In data science, root cause analysis helps businesses understand the semantics behind certain outcomes.


4. Define Confounding Variables.

A confounding variable is an external influence in an experiment. In simple words, these variables change the effect of a dependent and independent variable. A variable should satisfy below conditions to be a confounding variable :

Variables should be correlated to the independent variable.
Variables should be informally related to the dependent variable.
For example, if you are studying whether a lack of exercise has an effect on weight gain, then the lack of exercise is an independent variable and weight gain is a dependent variable. A confounder variable can be any other factor that has an effect on weight gain. Amount of food consumed, weather conditions etc. can be a confounding variable.