Introduction To data Science:

Data science has a intersection with AI but is not a subset of AI. it used some techique of AI,ML and deep learning and analysis data , visualizes then and built a model nad predict.

Data science is the art and science of drawing actionable insights from the data.

Applications: Retail,Bank, ecommerce,healthcare,tele communication and many more.

Data Science Flow Chart:

1.Data collection

2.Data preprocessing

3.Data analysis

4.Data visualization

5.Model Building

6.Deployment

Note: to bulid model (in step 5) , ML algo flow peocess as:

Model building flow chart

1.Importing data

2.Data Cleaning

3.Model building

4.Train model

5.Test model

6.improve eficiency ( if result is not accurate then , improve efficiency and after improving , prediction is done)

7.Prediction

 

Life Cycle of Machine Leaning

1.Define project objective

=> specify business problem

=>collect requirements

=>Decide project objective

2.Data Collection

=> Gather or collect the data which are needed for the prroject.

=> we can collect the data from two ways

i.primary data: collected by researches from first-hand source , collect by project team.

ii.Secondary data: collected by someone else and already benn passed through the statistical process.

Here yoyu collect the data but raw data might contain impurity like mising value,duplicated value e.t.c so datat are not in proper form or inconsistency of the data so we can't work on it. so we need data preprocessing.

3.Data Preprocessing

=>In this process we purify raw data or original data, we bring back to correct formate.

=>we use libraries like numpy,pandas,scikilearn e.t.c

=>There are many process in Data processing like

i.Data cleaning

  => filling missing data (NAn or ) is used if data is missing,row with no data are removed,remove irrelevant data)

ii.Data Transformation

  => Normalization(we used sckki learn here , normally we do data transforamtion to solve classificatiob problems)

iii.Dimensionality reduction

  => 3D data is converted into 2D (here we split data into dependent andn independent adn we give independent value to X and dependent value to Y and find relation between X and Y we perform data visualization

4.Data Visualization

=>Graphical represenrtation of data

=>data re transformation into table,graph,pi-chart,histogram e.t.c

=>libraries like matplotlibs are used in this process

=>After visualization, we create value from data,descover new patterns and findout useful insingh.

=>Here we compare the data and find the relation between them(i.e X and Y variable)

=> after fiinding relation between X and Y, now we need to select model

5.Model Selection

=> To identify which algorithms to use for a particular data

=>for example if you want to predict in numeric value, then you used linear regression, if calssification problem then used logistic regression,random forest or if you want to do clustering you might used k=means clustering.

6.Model Building

In Model building , we used Allgorithms to split data into Train and Test sets. like (70,30) or (80,20) i.e 70% of your data is used for trraining phase ands 30% to test the trrained model. and after that we we predict or get output from the model.

if model is not accurate then , model is trrained again and again untill the prediction is not accurated. So in thiis way model is improved according to desired output.

7.Deployment

Integrated ML model to existing environment to makke decission.