KNN - K Nearest neighbours

Machine Learning models make predictions from the past data available.

  • KNN is one of the simplest Supervised ML algorithm mostly used for Classification. It classifies based on how its neighbors are classified.
  • KNN stores available cases and classifies new cases based on similarity measure
  • Choice of K - sqrt(N)
  •  Lazy learning algorithm

How does (kids) teach kids to learn differentiate between cat & dog: 

  • type of claws, ear length, sound(bark vs meow), plays around vs not
  • kids identify given animal based on feature classification

Uses Cases:

  • Recommended systems- biggest use case in real-time
    • online shopping, OTT platforms, advertisement 
    • Content Search- documents having similar topics from billions of documents
    • Image & video recolonization
    • Height, weight -> derive T shirt size 
  • Predict dog category
  • Predict Over weight or not based on height & weight
  •  Predict Over Diabetes 
    • Pregnancies, Glucose, BP, Skin thickness, insulin, BMI, diabetes pedegree function, age
  • Sport liked based on age & Gender
  • Collaborative filtering & content based filtering
    • iphone with airpods
    • 2 users reading same article. recommend artiles 
    • Movies
      • Movie(movie id, title) & rating(user, movie,rating) details
      • create pivot matrix by user
      • using cosine similarity

Cosine Similarity & Cosine distance

  • Used in Recommendation systems
  • Cosine similarity is angle between two points.
    • cos-similarity = cos(theta) = angle between two points
    • cos-similarity ranges between -1 and 1

 Movie Recommendation system

  • Movie recommendation based on average Weighted value with simple correlation
    •  weighted rating = (mean rating for the movie * number of votes) + (mean vote across the whole report * min votes required to be listed in top 250)/(number of votes for movie +
      min votes required to be listed in top 250)
  •  Two types of recommendation systems
    • Content based filtering(based on type of content/movie)
      • This is based on content- movie summary using NLP tf-idf
      • Movies are recommended based on category of movie
      • eg: Netflix 
      • Action, Comedy, Romantic, Adventure,  
      • Geners, cast, crew, language, popularity, country, release_date, revenue,
    • Collaborative filtering
      • Recommendation is based on user behavior (like minded people behavior)
      • eg: Amazon site- recommendations to buy ear phones, headcover 
      • Collab multiple users based on behavior, pattern. 
      • Here it tries to behavior of user rather than type of movie

KNN-  Nearest neighbor item based on collaborative filtering

Pearson Correlation

Comments

Post a Comment

Popular posts from this blog

Statistics in Machine Learning

Note : In life cycle of data science more then 60% of time goes into data analysis like feature engineering, feature selection.  Feature engineering means Cleaning data, handle missing values, unbalanced data, category features.   Python packages Pandas - read dataset read_csv, head, isnull, getdummies, drop, concat, Numpy - work with arrays matplotlib.pyplot - for visualization Seaborn - for visualization heatmap, countplot, boxplot, Handling Categorical features one hot encoding for nominal variables label encoding for ordinal variables Ways for finding Outliers Scatter plot Box plot  z-score IQR Correlation :  Strength of association between two variables  This is both ways  A and B = B and A Regression If one of the variable is dependent & other is independent variable Regression equation = Average value of 'y' is a function of x R Square Significance of F & P values     Covariance(cov) Quantify relationship between features, rand...
Read more

Cluster Analysis

Clusters Homogenous within Heterogeneous across Based on field of interest There is no Objective function No dependent variable Popularly called subjective segmentation Segmentation develops on its own based on values of the input variables This is called Unsupervised learning How is Cluster different from decision tree techniques Decision tree technique requires a clearly defined objective funtion this is based on identifying independent and dependent variables. Why clustering Find groups which has elements similar to other groups elements of different groups are dissimilar to each other Generate treatment for specific group  Examples/Business Scenarios In Marketing Leading edge shoppers - looking for technological new products Big brand shoppers Seasonal shoppers Value conscious shoppers Fashion consumer groups Group driven by trends Group interested in Fashion Group purchasing when necessary Cluster analysis as part of business strategy...
Read more