# Category: Machine Learning

## 7 Steps of Machine Learning

The 7 steps of any Machine Learning problem to answering questions

1. Gathering Data
2. Preparing the Data
3. Choosing a Model
4. Training
5. Evaluation
6. Hyperparameter Tuning
7. Prediction

### Data Gathering

We will first gather data, in order to train our model we need data for example if we are predicting whether a drink is wine or beer, so we need features like colour and alcohol percentage.

### Data Preparation

We will randomise data, we can do Exploratory Data Analysis  that is to check biased that if we might have collected the beer data only that might result in beer biased data.

Data might need duplication, normalisation, error correction

Also to train the model we need to split the data in train & test, the test data will be used for model evaluation.

### Choosing a Model

We have lots of models created by researches over the years like some models works good with Image data, some are good at text based data. So we’ll try to choose a model according to our requirement.

### Training

Just like when someone is trying to drive a car, first the driver learns how to use brakes &  accelerator & over the time the drivers efficiency improves, the more he trains himself the more efficiency improves.

Y = mX+b

M-slope

B – y’s intercept

X – Input

Y – Output

So the values we can adjust are M & B only, there are lots of M in a model due to many features, so collection of M will be formed in to a matrix and denoted as W weight matrix and similarly for B we arrange the values into a matrix and it will be denoted as B Biases.

So in training we first initialise some random values to the model and try to predict the output with these values, So first the model performs vary poorly but after that we can compare it with the outputs that it should have produced and adjust the values in W & B then we will have more accurate predictions on the next time, each iteration (process of updating W & B) is called one training Step.

### Evaluation

In evaluation we test our model against the data which is never been used for training, this metric will allow us to see how model might perform against the data model has not seen yet. That how the model will perform in the real world

A good rule of thumb is to split the data in Training & Evaluation is 80%-20% or 70%-30%.

## 11 Data Mining Algorithms

1. Regression & Classification

• Linear
• Multivariate Linear
• Logistic
• Softmax
• Vectorization
• Optimizers and Objectives

2. Regularization

• Ridge regression

3. Clustering

• k – Means
• EM Algorithms

4. Unsupervised Learning

• Autoencoders
• PCA Whitening
• sparse coding

5. Neural Network

• Perceptrons
• Backpropagation
• Restricted Boltzmann Machines
• Learning Vector Quantization

6. Deep Learning

• Stacked Autoencoders
• Convolution Neural Networks (Feature Extraction, Pooling)
• Deep Boltzmann Machines
• Deep Belief Networks

7. Decision Trees

• ID3
• C4.5
• CART (Classification and regression tree)
• Random Forests

8. Bayesian

• Naïve Bayes
• Gaussian Naïve Bayes
• Bayesian Networks
• Conditional Random Fields
• Hidden Markov Models

9. Others

• Support Vector Machines
• Evolutionary Methods
• Reinforcement Learning
• Conditional Random Fields

10. Dimensionality Reduction

• PCA

11. Ensemble Methods

• Boosting
• Bagging

## Top Ten Places where AI and Machine Learning make our Life Easier

Creativity to make our surrounding automatic is our one and only aim left. Day by Day AI and Machine Learning automating more and more parts of our life.

We all have heard about AI thanks to movies for its introduction, but what about Machine Learning/ML. ML is the buzzword for most of us. Basically, ML makes computer to learn.

In a nut shell, ML is similar to our very first learning part of our childhood. We have a book containing a lot of pictures of fruits, animals, vegetables, and trees. These are teaching data set for any child. That data will be used to answer a question.For example, a picture is given to a child and he/she has to identify that pictures based on pictures saved in his/her mind. It is what the ML. ML continues to update its teaching data set based on correctly or incorrectly                                       credits:http://www.parlezwireless.com/
identification of things and get smarter and intelligent at completing its tasks over time. If you have used Google, Netflix, Amazon, Gmail, then you have interacted with machine learning (ML).

1. Recommendations
I am sure about recommendation type of thing if we use services like YouTube, Amazon or Netflix. Every click being monitored and recorded. Driven by Intelligent machine learning, these sites analyze our activity and compare it to the millions of other users to “recommend” or “suggest” other similar videos, products or films that we might like.
2. Online Search
AI is transforming Google and other search engine results by watching our response to result display. We click the results show on the very first page and we are done because we found what we are looking for. If not, then we go to the second page or refine our query at this point we assume that search engine didn’t understand what we want, so it learns its mistake and shows the better result in the near future.
3. In Hospitals
credits: http://assets.fastcompany.com
Due to its nature of analyzing vast amounts of data, ML takes the first place to process information and spot more pattern like cancer or eye diseases than a human can by several orders of magnitude.Computer-aided diagnosis (CAD) can help radiologists find early-stage breast cancers that might otherwise be missed, and it can identify 52% of these missed cancers roughly a year before they were actually detected. Zebra Medical Systems is an Israeli company that applies advanced machine learning techniques to the field of radiology. It has amassed a huge training set of medical images along with categorization technology that will allow computers to predict multiple diseases with better-than-human accuracy. In 2016, the company unveiled two new software algorithms to help predict, and even prevent, cardiovascular events such as heart attacks.
4. Data Security
According to Kaspersky, between January and September, 2016 ransomware attacks on business increased from once every 2 minutes to once every 40 seconds. Symantec also reported high levels of ransomware attacks, over 50,000 in March 2016 alone. A report by Osterman Research indicates 47% of organizations in the US in 2016 had been targeted at least once. A survey in the UK suggested 54% of businesses had been attacked at least once. Friday, May 12, 2017, saw one of the largest most widespread attacks to date – the WannaCry ransomware. According to Deep Instinct new malware tends to have almost the same code as the previous one only 2 to 10% changes. Due to the slight change in code ML can predict which files are malware or not with great accuracy.
5. Email spam filtering
According to Computer World magazine, the average employee gets 13 spam messages a day – and over 80 percent of all the email messages zipping around the Internet are spam. Microsoft founder Bill Gates is the most spammed man in the world, with 4m emails arriving in his inbox each day. All credit goes to ML which filter all emails and classify them into spam and not spam.                                                          credits:http://www.asistiletisim.com
6. Marketing Personalization
Personalized marketing is the ultimate form of targeted marketing. To sell more we have to serve better and to serve better we have to understand customers. This is the base idea behind marketing personalization. Companies can personalize customer emails, which products will show up as recommended, offer they see, coupons and so on, these are just the tip of the iceberg. All above things are achieved by the advance ML algorithm.
7. Fraud Detection
ML and AI are used and become better day by day at spotting potential cases of fraud or anomaly detection across many different fields. The Royal Bank of Scotland (RBS) for example, is using machine learning to fight money laundering. Companies have a lot of data and they use ML to compare millions of transactions and can precisely distinguish between legitimate and fraudulent transactions between buyers and sellers.
8. Natural Language Processing (NLP)
Virtual personal assistants – likes of Siri, Alexa, Cortana and Google Assistant – are able to follow instructions because of voice recognition and it is NLP. NLP process human speech and match it to best-desired command and respond it in a natural way.