Learnings from Week 1 of MLDL show .

Hello , Week 1 of The #MLDLS2021 is Completed and I will be sharing my learnings from it .

                                                 **Learnings from First Lecture**

1. Intro to ML and it's sub branches
2. Intro to ML algorithms: supervised, un- supervised, regression, classification with examples .
3. We were divided into breakout rooms for discussing applications of ML .
4. Every algorithm was explained well with examples like when should you use regression and when classification .
   • Ex: For predicting rain(i.e. will it rain tomorrow or not) we will use classification and not regression.
   • Ex: When you have data which seems to be divided into groups ,you will use clustering algorithm as it will divide the data into groups .
5. A quiz on these topics(I scored a 10 (Hurray)) .

                                                  **Learnings of Second Lecture** 

1. Linear Regression :
   • Simply, a graph line is plotted by the algorithm according to the x and y values provided.
   • If there are multiple input variable (Price of house vs. area+rooms+color): Multiple Regression
2. Loss functions:
   • The difference of actual and predicted value in a graph and their average is the value of Loss function .
   • The less loss value ,the better algorithm .
   • If the loss value is higher try for searching another algorithm.
   • The more data and training of model ,the better .
   • Ideal training and testing of data should be 8:2 or 7:3 .

3. Important libraries for machine learning :

   • Numpy : NumPy basically provides n-dimensional array object. NumPy also provides mathematical functions which can be used in many calculations.
   • Code: import numpy as np

              arr = np.array([[1,2,3],[4,5,6]])
              print("Numpy array\n {}".format(arr))
     

   • Matplotlib : Matplotlib is scientific plotting library usually required to visualize data. Importantly visualization is required to analyze the data. You can plot histograms, scatter graphs, lines etc.(In sort for plotting graphs ) .

   • Code: import matplotlib.pyplot as plt

            x = [1,2,3]
            y = [4,5,6]
            plt.scatter(x,y)
            plt.show()
     

   • Pandas : Pandas is used for data analysis it can take multi-dimensional arrays as input and produce charts/graphs.
   • Code: import pandas as pd

            age = {'age': [4, 6, 8, 34, 5, 30, 41] }
            dataframe = pd.DataFrame(age)
            print("all age:\n{}".format(dataframe))
            filtered = dataframe[dataframe.age > 20]
            print("age above 20:\n{}".format(filtered))
     

   • sklearn : Sklearn is used for creating the linear regression model Refer official doc : https://scikit-learn.org/stable/modules/generated/sklearn.linear_model.LinearRegression.html
4. Dataset websites :
   • Kaggle : kaggle.com
   • Google data search : datasetsearch.research.google.com
   • UCI machine learning : archive.ics.uci.edu/ml/index.php
5. Platform for making a machine learning model :
   • Google Colab: colab.research.google.com/?utm_source=scs-i..
   • It gives you free to use GPU for better processing .
6. A quiz for testing our knowledge( got 9 on 10) .

The best part about the presentation is that we get to see Machine learning memes after every topics ,and we love it .

Some memes for you too :

Next step : FROM MLDL Show.

All thanks to our Awesome mentors : Rucha,Vedant and Harsh.