Churn Prediction Analysis Using Machine LearningLast Updated on May 3, 2021
Churn means customers or users who left the services or migrates to the competitor in the industry. It is very important for any organization to keep its existing customer and attract new ones if one of them fails it is bad for business. The goal is to explore the possibility of machine learning for churn prediction to retain a competitive edge in the industry. It is a classification problem so the goal is to predict weather user left the services or not.
The project is structured as follows:
- Data cleaning
- Exploratory Data Analysis
- Data Preprocessing
- Oversampling Technique
- Model Creation and Evaluation
- Improving the Model
- Model save
In order to measure the performance of the model, the Area Under Curve (AUC) standard measure is adopted i.e. 0.91 and the AUC value obtained. We have achieved an overall accuracy of almost 80% with just direct implementation of the model without performing hyperparameter tuning. And by applying hyperparameter tuning we get accuracy approx 85% and improve the model.
Share with someone who needs it
TwosidednewsLast Updated on May 3, 2021
Juxtaposition of sources from opposite sides of the political spectrum.
The primary inspiration of our project is the growing filter bubbles in our country and in our world. The ability of people to only see posts on social media and news sites that agree with their point of view is a worrying development of the 21st century that we are trying to combat. Our project being two-fold (Chrome Extension and website) means that we can both provide an in depth overview of an issue for those who are actively curious about learning more about an issue through our website, and also passively prompt users with alternative interpretations of news stories for users who are not consistently conscious of filter bubbles.
What it does
Two Sided News attempts to give two views on any story: liberal and conservative. By searching keywords to a topic the user would like to read about, our website displays articles side by side. The user can then choose to read whichever perspective they please, or both, and come up with their own interpretations of the story. The chrome extension version allows you to directly look at another article from the opposite view. This way, users can continue to browse articles on websites that they are comfortable with, but have the option to read the another side of the story through this extension.
How we built it
Challenges we ran into
We all had to learn how to use Github for group version control, and had to teach each other web development basics. In addition, our collective experience with Chrome extensions was that one of our members had attended the Chrome extension [email protected] workshop. Yet we still managed to make our thing!
Accomplishments that we're proud of
We are extremely happy to have been able to complete this project in less than a day. Our website lets users read online news articles, but we are especially proud that our product can allow them to become more informed about other perspectives. The portable version of our product, the Chrome extension, can even be used on the go for readers who don't use our website.
What we learned
What's next for TwoSidedNews
Future features would be to upgrade the news search. One way to do this is implement Machine Learning that will take in a database of articles and learn which articles are liberal or conservative leaning. This way, the article returned from a query would not be restricted to a particular news outlet. For example, if one news outlet has both left and right leanings, the same news outlet can be displayed on both sides. Another feature would be a website format that can be its own sustainable news website. This would include tags like "Popular" or "Health" where the trending stories for each would be displayed instead of having to specifically querying for an article/subject. This would also include updating the design of the website to best accommodate users' experience.
Try it out
Visit the website here
Check out the Chrome Extension in the Google Chrome Store here
Human Computer Interaction Using Iris,Head And Eye DetectionLast Updated on May 3, 2021
HCI stands for the human computer interaction which means the interaction between the humans and the computer.
We need to improve it because then only it would improve the user interaction and usability. A richer design would encourage users and a poor design would keep the users at bay.
We also need to design for different categories of people having different age,color,gender etc. We need to make them accessible to older people.
It is our moral responsibility to make it accessible to disabled people.
So this project tracks our head ,eye and iris to detect the eye movement by using the viola Jones algorithm.But this algorithm does not work with our masks on as it calculated the facial features to calculate the distance.
It uses the eucledian distance to calculate the distance between the previous frame and the next frame and actually plots a graph.
It also uses the formula theta equals tan inverse of b/a to calculate the deviation.
Here we are using ANN algorithm because ANN can work with incomplete data. Here we are using constructive or generative neural networks which means it starts capturing our individual images at the beginning to create our individual patterns and track the eye.
Here we actually build the neural network and train it to predict
Finally we convert it to mouse direction and clicks and double clicks on icons and the virtual keyboard.
As a contributing or moral individuals it is our duty to make devices compatible with all age groups and differently abled persons.
Quiz App In AndroidLast Updated on May 3, 2021
§ The “QuizApp” has been developed to override the problems prevailing in the practicing manual system
§ This App is supposed to eliminate and in some cases reduce the hardships faced by this existing system. Today internet become reality and usage of internet become very much popular and there is tremendous increase of internet in all over the world for education purpose.
§ The QuizApp provides complete functionality of evaluation and assessing student’s performance skills. Quiz Application can lead to error free, secure, reliable and fast.
§ The Quizzes will form the backbone of the automated process and will play an important role in generation of unique sets of questions.
The Quiz application is used for conducting quiz for students or this software can also be used by the company for the recruitment process.
At first, the student is needed to register his/her name along with all the information needed and need to select username and password for the login process. Using this username and password, the student can login into the “QuizApp” App.
Next procedure is answering the quiz. As soon as the student selects the Category and set, the questions with 4 options will be displayed.
The student has to select any one option and click on next option. This will continue till the end of the question. At the end final result will displayed to user.
Student has facility to bookmark question for further reference.
- Immediate Results
- Category wise Question Set
- Authentication using Firebase
Telecom Churn PredictionLast Updated on May 3, 2021
Business Problem Overview In the telecom industry, customers are able to choose from multiple service providers and actively switch from one operator to another. In this highly competitive market, the telecommunications industry experiences an average of 15-25% annual churn rate. Given the fact that it costs 5-10 times more to acquire a new customer than to retain an existing one, customer retention has now become even more important than customer acquisition. For many incumbent operators, retaining high profitable customers is the number one business goal. To reduce customer churn, telecom companies need to predict which customers are at high risk of churn. In this project, you will analyse customer-level data of a leading telecom firm, build predictive models to identify customers at high risk of churn and identify the main indicators of churn.
Modelling Build models to predict churn. The predictive model that you’re going to build will serve two purposes:
1)It will be used to predict whether a high-value customer will churn or not, in near future (i.e. churn phase). By knowing this, the company can take action steps such as providing special plans, discounts on recharge etc.
2)It will be used to identify important variables that are strong predictors of churn. These variables may also indicate why customers choose to switch to other networks.
Multi Label Question Classification For Agricultural DomainLast Updated on May 3, 2021
MULTI LABEL QUESTION CLASSIFICATION
Basically, Multi-Label means, the questions asked by a client are of what type?
Here our ML model should be able to segregate the questions into descriptive type question or one-word type question or both and then answer it effectively.
We have classified the data into three major types:
1) Definition Type Question 2) Descriptive Type Questions 3) Factoid Type Questions
KEYWORD IDENTIFIER Keywords are used to identify the question type.
The input query is scanned and the primary keywords such as who, when etc. are identified.
These keywords help in finding out the expected answer type. But words like how and which do not give a clear idea about the question type.
To get a clear idea about the question type and obtain the relevant answer, additional keywords are required.
These are known as secondary keywords. The secondary keywords provide additional information about the question type which further helps in extracting the answer from the document.
a)Which sector is the backbone of the Indian economy? Which is the best season for a particular crop?
Here the “which” keyword acts as the primary key. This helps us understand the question type as which and the expected answer for such type of question is a factoid type of answer which precisely answers the type of “sector”
Data preprocessing techniques such as data redundancy, stop words, data cleaning, puntuactions, etc are done first.
Using the Naive Bayes Classifier &Python's Scikit-learn package we were able to upload the Model database containing 50 questions with their answers.
Naive Bayes and SVM is a simple technique for constructing classifiers: models that assign class labels to problem instances, represented as vectors of feature values, where the class labels are drawn from some finite set.
Training and testing of data are done and then the model predicts the appropriate answer.