Scribd
Upload
21 views17 pages

07 - Model Selection & Building

The document discusses natural language processing and text preprocessing. It covers selecting and building machine learning models using algorithms like random forest. Random forest creates multiple decision trees and aggregates their predictions to improve accuracy. The document provides examples of how random forest works in areas like online shopping, finance, and selecting the best machine learning algorithm for a project.

Uploaded by

Omar Ben
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
21 views17 pages

07 - Model Selection & Building

The document discusses natural language processing and text preprocessing. It covers selecting and building machine learning models using algorithms like random forest. Random forest creates multiple decision trees and aggregates their predictions to improve accuracy. The document provides examples of how random forest works in areas like online shopping, finance, and selecting the best machine learning algorithm for a project.

Uploaded by

Omar Ben
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 17

‫معالجة اللغات الطبيعية‬

Natural Language Processing(NLP)

Ahmad Shhadeh
٥– ‫تهيئة البيانات النصيه‬-‫تطبيق عملي‬
SMS SPAM Filtering
‫إختيار و بناء النموذج –خوارزمية تعلم االله‬
Model Selection & Building

AHMAD SHHADEH
Data Science Project life cycle

 Business Understanding
 Data Collection
 Data Preparation
 Exploratory data analytics(EDA)
 Model Building
 Model Evaluation
 Model Deployment
Ahmad Shhadeh
ML algorithm selection

➢ There are so many Machine learning in the world ,Which machine learning Should we use ? There is no straightforward and
sure-shot way to choose the right MLA. Determining which algorithm to use depends on many factors like:
➢ The problem statement ,The kind of output we are looking .
➢ Type and size of the data,
➢ The available computational time/Resources(Memory, Type of processors ),
➢ Number of features, and observations in the data
➢ …etc
➢ Key skills can help :
➢ Machine Learning Types : Supervised and Unsupervised
➢ Domain knowledge to filter down (CV, NLP, anomaly detection .. etc )
➢ Data science project Pipeline
Types of machine learning
Types of machine learning

https://scikit-learn.org/stable/tutorial/machine_learning_map/
Source : https://blogs.sas.com/content/subconsciousmusings/2020/12/09/machine-learning-algorithm-use/
Introduction to Ensemble Learning

 Ensemble :A technique that create multiple models and then combine them to produce a
better result
 Why learn one classifier when you can learn many
 Example: I have created a short movie about machine learning and need to get a feedback before
making it public
 1st Model: asking two of my friends
 2nd Model: asking 5 colleagues on the machine learning domain
 3rd Model :creating a small survey and get feed back from 20 people
 The responses, would be more generalized and diversified since we have people with different skill set
and different relationships ,This is a better approach to get honest ratings
 With these examples, you can infer that a diverse group of people are likely to make better decisions as
compared to individuals.
Random Forest in real life

❖ You want to purchase a new car!!!!!!


 will you walk up to the first car shop and purchase one based on the advice of the dealer? OR
 You would browser a few web sites/mobile App (haraj ,souq..etc)
 check the posted reviews
 Compare different car models, checking for their features and prices.
 You will also ask some of your friends for their opinion.
In Summary , you wouldn’t directly reach a conclusion, but will instead make a decision considering the
opinions of other people as well.

❖ Who want to be millionaires?


 Asking the audience option
Random Forest in real life

 In the world of finance and investments,


▪ The basic concepts of the investments is to build a bunch of
uncorrelated models,
▪ Each with a positive expected return,
▪ then put them together in a portfolio to earn massive alpha (alpha =
market beating returns).

 Last example for online shopping


 We rely on multiple sources (never trust a solitary Amazon
review), and therefore, not only is a decision tree intuitive,
but so is the idea of combining them in a random forest.

In short : A Multiple number of relatively uncorrelated models (trees) operating as a committee will
outperform any of the individual constituent models.
What is Random Forest?

 Random forest is an ensemble learning method that constructs a collection of


decision trees and then aggregates the predictions of each tree to determine
the final prediction.
 A decision tree is the building block of a random forest and is an intuitive
model.
 Combining the weak models which are produced by individual decision trees
to get a strong model
 In general, the higher the number of trees in the forest gives the high
accuracy results.
Random Forest Characteristics
 Advantage :
 Very versatile and powerful machine learning algorithm.
 Can solve both type of problems i.e. classification and regression, Accepts various types of inputs as well, may it
be ordinal or continuous data.
 Easily handles outliers, missing values, skewed data, the data doesn't even have to be on the same scale.
 Less likely to overfit than some of the other machine learning models.
 Providing the feature importance by identifying the most significant variables so it can use for dimensionality
reduction methods.
 Disadvantage:
 It surely does a good job at classification but not as good as for regression problem
 Random Forest can feel like a black box approach for statistical modelers - you have very little control on what
the model does.
Random Forest for our project(SMS Spam Filtering )

 For our project :SMS Spam filtering


 We can build a random forest with 100 decision trees in it.
 Then each decision trees are built independently, and it will predict either spam or ham.
 Assume 70 of those decision trees predict spam and 30 predict ham.
 Then will apply simple voting method for the trees.
 Max Voting
 Averaging
 Weighted Averaging

 Then the final prediction of the random forest model will be spam if we applied the Max Voting method.
 In our project will use the simplest voting technique which is Max Voting
Max voting method

 The max voting method is generally used for classification


problems.
 In this technique, multiple models are used to make predictions
for each data point. The predictions by each model are
considered as a ‘vote’.
 The predictions which we get from the majority of the models
are used as the final prediction.
 For example, when you asked 5 of your colleagues to rate your
movie (out of 5); we’ll assume three of them rated it as 4 while
two of them gave it a 5. Since the majority gave a rating of 4,
the final rating will be taken as 4.
sklearn.ensemble.RandomForestClassifier
 https://scikit-learn.org/stable/modules/generated/sklearn.ensemble.RandomForestClassifier.html
Parameter/Attribute/Met Descriptrion
hod

n_estimators (par) The number of trees in the forest., default=100.


n_jobs, (par) default=None ,The number of jobs to run in parallel. fit, predict, decision_path and apply are all
parallelized over the trees. None means 1 unless in a joblib.parallel backend context. -1 means
using all processors

fit(X_train, y_train) (Method) Build a forest of trees from the training set (X, y).

predict(X)test) (Method) Predict class for X.


Feature Importances indicate what predictor variables the random forest considers most important,it can be used for
(Attributes) feature engineering by building additional features from the most important. We can also use
feature importances for feature selection by removing low importance features.
Split Training and Testing Data Sets

 Machine learning is training an algorithm on a set of known examples with a clear goal of generalizing to unseen
examples.
 The train-test split is a technique for evaluating the performance of a machine learning algorithm, It can be used
for classification or regression problems and can be used for any supervised learning algorithm.
 The procedure involves taking a dataset and dividing it into two subsets.
• Train Dataset: Used to fit the machine learning model.
• Test Dataset: Used to evaluate the fit machine learning model: not used to train the model
 The objective is to estimate the performance of the machine learning model on new data: data not used to train the
model.
Thank You

You might also like