Top 7 Algorithms to Know for Building Recommender Systems

Want to learn how to build awesome rec systems? Start here:

1. Item-based collaborative filtering - https://lnkd.in/gNkK9HP

2. Non-negative matrix factorization - https://lnkd.in/giUXS-E

3. Contenting-based filtering - https://lnkd.in/gFvacKs

4. kNN - https://lnkd.in/gUvEqsR

5. Knowledge-based rec systems - https://lnkd.in/gW5muUV

6. Clustering - https://lnkd.in/gdTkW3K

7. Vector similarity measures: Pearson, Jaccard, cosine - https://lnkd.in/gn55WhP - https://lnkd.in/g5iuCPF - https://lnkd.in/gEMj9hp

Bonus 1: Bayesian networks - https://lnkd.in/gKF2Y87
Bonus 2: Hidden Markov models - https://lnkd.in/gzzNGtj

Start by getting familiar with collaborative filtering at a high level - https://lnkd.in/gtE5HRB

Then grab a dataset:

* Last.fm - https://lnkd.in/gUr-8U6
* MovieLens - https://lnkd.in/gNv4FYN
* Others - https://lnkd.in/gnqu7XR

Next, start exploring the algorithms and experimenting with them on your data.


Get familiar with these 7 concepts and you'll be ready to take on almost any recommendation problem in no time.

✴️ @AI_Python_EN

Apparently saving as JPG can reverse the perturbations back to the original image.

"A study of the effect of JPG compression on adversarial images"
https://arxiv.org/pdf/1608.00853.pdf

Here is an excerpt from the conclusion:

Our experiments demonstrate that JPG compression can reverse small adversarial perturbations
created by the Fast-Gradient-Sign method. However, if the adversarial perturbations are larger, JPG
compression does not reverse the adversarial perturbation.

✴️ @AI_Python_EN

A Pattern-Based Method for Medical Entity Recognition From Chinese Diagnostic Imaging Text

The identification of medical entities and relations from electronic medical records is a fundamental research issue for medical informatics. However, the task of extracting valuable knowledge from these records is challenging due to its high complexity.

The method proves to be stable and robust with different amounts of testing data. It achieves a comparatively high performance in the CHIP 2018 open challenge, demonstrating its effectiveness in extracting tumor-related entities from Chinese diagnostic imaging text.

Paper: https://lnkd.in/g3uezmR
#LSTM #RNN #deeplearning #healthcare #algorithm

✴️ @AI_Python_EN

NLP is the most requested topic that rarely I cover, please select one from this 14 you need go more detail?

1. Machine Translation
(https://lnkd.in/fAYvEne)
2. Question Answering (Like Chat-bot)
(https://lnkd.in/fFZmP4f)
3. Sentiment Analysis
(https://lnkd.in/fUDGAQW)
4. Text Search (with Synonyms)
(https://lnkd.in/fnU_a_H)
5. Text Classifications
(https://lnkd.in/f8mjKAP)
6. Spelling Corrector
(https://lnkd.in/f8JXNUv)
7. Entity (Person, Place, or Brand) Recognition
(https://lnkd.in/f2fzgAa)
8. Text Summarization
(https://lnkd.in/fdzWqXC)
9. Text Similarity
(https://lnkd.in/fv_sWuM)
10. Topic Detection
(https://lnkd.in/fxmhJZc)
11. Emotion Recognition
(https://lnkd.in/fK4m66Q)
12. Language Identification
(https://lnkd.in/fqfjxF9)
13. Document Ranking (https://lnkd.in/fJZnkqz)
14. Fake News Detection
(https://lnkd.in/fkrkF8Q)

✴️ @AI_Python_EN

Here is a #MachineLearning math quiz for you.

There are several loss functions (0-1, logarithmic, quadratic, exponential etc) and there is a risk function as well.

Can you define your own loss function?
#mathematics #deeplearning
#ai
✴️ @AI_Python_EN

From time to time I still hear the comment that, since we now have AI, statistics isn't needed anymore.

First, true AI - Artificial General Intelligence - has not yet arrived. "Machines," i.e., software, cannot truly "think" apart from performing specific tasks and solving problems they've been programmed to solve. They can recognize patterns in our behavior but cannot feel emotion.

Task-specific AI is now everywhere around us and having an ever-greater impact on our daily lives and jobs. Statistical methods are sometimes part of the engines of these AIs, and some familiar statistical techniques which have been automated have been disingenuously branded as AI.

Text mining, voice and image analysis, robot navigation are examples of tasks that require specialized software, whether or not we call it AI, machine learning or just software. Again, in some cases, statistics is part of these programs.

However, when we need to design research and analyze data for particular purposes, statistics is still essential. It may be augmented by task-specific AI, but cannot be replaced by it. AI and statistics are sometimes direct competitors but more typically synergize or address different objectives.

I sense some confusion about this and hope this short post will help clear up some of it.

I hear many strange comments about "AI." For example, credit scoring is at least 60 years old. Originally, a human would score a loan application based on a pre-determined formula with adding machines. There was no human discretion, though other factors also had an impact on whether or not the loan was approved. Likewise, insurance has been priced according to pre-determined formulas for many decades. Automating these calculations does not change the basic principles. Many bureaucratic decisions are also determined by rules that will not bend, and this was once seen as progress.

If you'd like to learn more about #AI, "AI For Ordinary Folks" may be of help:
https://greenbookblog.org/2019/03/27/ai-for-ordinary-folks/

✴️ @AI_Python_EN

We’ve now released the 0.2.0 version of the Python StanfordNLP library! Highlights:
1) Much smaller model sizes (9x smaller for English, 11x for German, ...);
2) Substantial speedup of the neural lemmatizer;
3) Easier-to-use CoreNLP interface. More at:
https://github.com/stanfordnlp/stanfordnlp/releases/tag/v0.2.0

✴️ @AI_Python_EN

Generating Game of Thrones Characters Using StyleGAN
article: https://blog.nanonets.com/stylegan-got/

gitHub repo: https://github.com/iyaja/stylegan-encoder
#deeplearning #ArtificialIntelligence
✴️ @AI_Python_EN

There's increasing interest in time-series analysis in the #DataScience community. I see this as a very positive development, as TSA has seldom gotten the attention from data scientists I feel it merits.

Econometricians and statisticians in other fields such as meteorology have been analyzing time-series data for many decades. Before diving into LSTMs and other machine learning tools - which statisticians also use - I think it would be wise to read up on other methods statisticians employ and what they've learned from more than a century of experience.

Here are some good books on TSA:

- Time Series Analysis (Wei)
- Time Series Analysis and Its Applications (Shumway and Stoffer)
- Forecasting (Hyndman and Athanasopoulos)
- Multiple Time-Series Analysis (Lütkepohl)
- Time Series Analysis by State Space Methods (Durbin and Koopman)
- Time Series Modelling with Unobserved Components (Pelagatti)
- Hidden Markov Models for Time Series (Zucchini)
- GARCH Models (Francq and Zakoïan)
- Nonparametric Econometrics (Li and Racine)

There are academic journals as well. (I subscribed to the Journal of Forecasting and the Journal of Time Series Analysis until they jacked up their subscription fees.)

TSA is a difficult topic, IMO, but more important than ever in the age of big data.
#book
✴️ @AI_Python_EN

Multivariate Data Visualization with python


9 Visualization that useful and doable to any tabular dataset

Try these 9 VisualizationCreated by :Rahul Agarwal

#machinelearning #artificialintelligence #datascience #ml #ai #deeplearning #python #visualization #visualisation

✴️ @AI_Python_EN

Data Visualization is a very important step in Data Science, so we should try to MASTER it.

Here are the useful links for Data Visualization -

1)Quick and Easy Data Visualizations in Python with Code.
(https://lnkd.in/fXJ-_Y8)

2)10 Useful Python Data Visualization Libraries for Any Discipline.
(https://lnkd.in/fBxbHwr)

3)Top 50 matplotlib Visualizations – The Master Plots (with full python code).
(https://lnkd.in/fGrnGax)

4)Data Visualization Effectiveness Profile.
(https://lnkd.in/f3v52Fd)

5)The Visual Perception of Variation in Data Displays.
(https://lnkd.in/fm-TbPM)

6)Matplotlib Tutorial – A Complete Guide to Python Plot w/ Examples.
(https://lnkd.in/fFkUgQP)

7)Interactive Data Visualization in Python With Bokeh.
(https://lnkd.in/fEfQAvg)

8) Data Visualization in R
https://lnkd.in/fEvZB_N

9) The Next Level of Data Visualization in Python (Plotly)
https://lnkd.in/fKn4cPM

#datascience #dataanalysis #datavisualization #python #r

✴️ @AI_Python_EN

How To Define A Convolutional Layer In PyTorch

http://bit.ly/2JFFlxI
#AI #DeepLearning #MachineLearning
#DataScience


✴️ @AI_Python_EN

📍list of awesome papers for electronic health records(EHR) mining, machine learning, and deep learning


🔗 Repo github: https://github.com/hurcy/awesome-ehr-deeplearning

@AI_Python_EN

Some python packages that make #deeplearning easier


✴️ @AI_Python_EN

A Step-by-Step Introduction to Object Detection Algorithms (10 pages) - Pulkit Sharma

#machinelearning #artificialintelligence #datascience #ml #ai #deeplearning #objectdetection

✴️ @AI_Python_EN

#DataScience Learning Path For Complete Beginners:
https://bit.ly/2JhcjXW
#BigData #MachineLearning #AI #DataScientists #Python

✴️ @AI_Python

Decision Tree Classifier from Scratch
http://bit.ly/2HAW0lh

#MachineLearning

✴️ @AI_Python

✴️ @AI_Python

A Practical Introduction to #DeepLearning with #Python
http://bit.ly/2QbmApb
#MachineLearning
✴️ @AI_Python

There are now many methods we can use when our dependent variable is not continuous. SVM, XGBoost and Random Forests are some popular ones.

There are also "traditional" methods, such as Logistic Regression. These usually scale well and, when used properly, are competitive in terms of predictive accuracy.

They are probabilistic models, which gives them additional flexibility. They also are often easier to interpret, critical when the goal is explanation, not just prediction.

They can be more work, however, and are probably easier to misuse than newer methods such as Random Forests. Here are some excellent books on these methods that may be of interest:

- Categorical Data Analysis (Agresti)
- Analyzing Categorical Data (Simonoff)
- Regression Models for Categorical Dependent Variables (Long and Freese)
- Generalized Linear Models and Extensions (Hardin and Hilbe)
- Regression Modeling Strategies (Harrell)
- Applied Logistic Regression (Hosmer and Lemeshow)
- Logistic Regression Models (Hilbe)
- Analysis of Ordinal Categorical Data (Agresti)
- Applied Ordinal Logistic Regression (Liu)
- Modeling Count Data (Hilbe)
- Negative Binomial Regression (Hilbe)
- Handbook of Survival Analysis (Klein et al.)
- Survival Analysis: A Self-Learning Text (Kleinbaum and Klein)

#statistics #book #Machinelearning

✴️ @AI_Python