Machine Learning for Dummies

A DS/ML enthusiast

[Machine Learning for Dummies @DummiesMachine]

Possibly the best explanation to #PCA and #LDA ! Link below -

https://t.co/SsLYQ6EpM2

[Machine Learning for Dummies @DummiesMachine]

In #Gaussian Naive Bayes, continuous values associated with each feature are assumed to be distributed according to a #Gaussian distribution. #bellshapedcurve #sameasnormaldistribution

[Machine Learning for Dummies @DummiesMachine]

In #Gaussian Naive Bayes, continuous values associated with each feature are assumed to be distributed according to a #Gaussian distribution. #bellshapedcurve #sameasnormaldistribution

[Machine Learning for Dummies @DummiesMachine]

Gradient boosting machines are generally very slow in implementation because of sequential model training.Thus, #XGBoost is focused on computational speed and model performance.

[Machine Learning for Dummies @DummiesMachine]

#LightGBM beats all the other algorithms when the dataset is extremely large. LightGBM is a gradient boosting framework that uses tree-based algorithms and follows #leafwise approach while other algorithms work in a #levelwise approach pattern.

[Machine Learning for Dummies @DummiesMachine]

Since #DecisionTrees, #Random Forests, #XGBoost takes care of the missing values itself, you do not have to impute the missing values.

[Machine Learning for Dummies @DummiesMachine]

#Boosting in general decreases the bias error and builds strong predictive models. #Boosting has shown better predictive accuracy than #bagging, but it also tends to over-fit the training data as well #ensemblemodelling

[Machine Learning for Dummies @DummiesMachine]

#Earlystopping is when you are training your dataset but do not go all the way till the point you cost function is the least so as to keep the weights mid sized.The more you train,the bigger the weights of your model become hence it simulateneously gives #regularization effect

[Machine Learning for Dummies @DummiesMachine]

A bit about regualrization techniques - L2 will reduce your parameter weights a bit thus diminishing effect of 1 parameter too much while L1 totally removes some parameters by adjusting its wight to 0 #machinelearning #datascience #statistics #modellingtechniques #AI #regression

[Machine Learning for Dummies @DummiesMachine]

#Kernel function in #SVM takes low dimensional input space and transform it to a higher dimensional space #machinelearning #datascience #statistics #modellingtechniques #AI #regression #classification #MLalgorithms

[Machine Learning for Dummies @DummiesMachine]

#SVM has a feature to ignore outliers and find the hyper-plane that has maximum margin. #machinelearning #datascience #statistics #modellingtechniques #AI #regression #classification #MLalgorithms

[Machine Learning for Dummies @DummiesMachine]

SVM selects the hyper-plane which classifies the classes accurately PRIOR to maximizing margin. #machinelearning #datascience #statistics #modellingtechniques #AI #regression #classification #MLalgorithms

[Machine Learning for Dummies @DummiesMachine]

A median is not affected by outliers. So when you see your dataset having too many outliers, use #median instead #mean for #binning purposes.

[Machine Learning for Dummies @DummiesMachine]

There are various methods to calculate the similarity between two objects while building a #recommendationsystem. Distance scores like #euclidean, #cosine, #pearson #corelation can be used, there is no good or bad. decide using all three and see which gives better results

[Machine Learning for Dummies @DummiesMachine]

#collaborative methods are largely #memorybased and #modelbased. Memory based models are based on user-user or item-item frameworks while model based rely on methods like #matrixfactorization in order to deal with sparse datasets#machinelearning #datascience #statistics

[Machine Learning for Dummies @DummiesMachine]

#Recommendor systems are widely divided into #contentbased and #collaborativefiltering. For former, you are aware of different features vectors that define the product to be recommended while in latter, it is an incremental process of learning product as well as user features

[Machine Learning for Dummies @DummiesMachine]

Whenever you see terms like ||u|| , it basically means length of #vector u which is sq. root of sum of its x and y projection #machinelearning #datascience #statistics #modellingtechniques #AI #regression #classification #MLalgorithms

[Machine Learning for Dummies @DummiesMachine]

A #vector is nothing but a nX1 #matrix ! #machinelearning #datascience #statistics #modellingtechniques #AI #regression #classification #MLalgorithms

[Machine Learning for Dummies @DummiesMachine]

While #sigmoid activation is used for binary classification tasks, #softmax activation is used for Used for the multi-classification tasks #machinelearning #datascience #statistics #modellingtechniques #AI #regression #classification #MLalgorithms

[Machine Learning for Dummies @DummiesMachine]

Some info on activation functions. #sigmoid to convert outputs between 0 to 1, #tanh for -1 to 1, #relu where anything less thn zero has output of 0, #softmax again to collapse outputs between 0 to 1 - used in #multilable #classification