aayushs879/Text-Classification-with-a-Multi-Lingual-Corpus

Stars
2
Language
Python
Created almost 6 years ago
Updated almost 6 years ago

aayushs879/Text-Classification-with-a-Multi-Lingual-Corpus

aayushs879

There are no reviews yet. Be the first to send feedback to the community and the maintainers!

Societe Generale (SocGen) is a French multinational banking and financial services company. With over 1,54,000 employees, based in 76 countries, they handle over 32 million clients throughout the world on a daily basis. They provide services like retail banking, corporate and investment banking, asset management, portfolio management, insurance and other financial services. While handling customer complaints, it is hard to track the status of the complaint. To automate this process, SocGen wants you to build a model that can automatically predict the complaint status (how the complaint was resolved) based on the complaint submitted by the consumer and other related meta-data.

Faster-RCNN-on-BCCD-Dataset

This project is an implementation of Faster CNN on BCCD Dataset of bloodcells. We need to detect wheter the cells in a picture are RBCs WBCs or Platelets. I used tensorflow object detection api for this.

Tripadvisor-Review-dataset

This was the review dataset from tripadvisor of 21 hotels in The Strip, Las Vegas. Please go through the notebook provided

CNN-based-vegetable-recognition-model

This project was a part of the work done by me and my team for All India Research Championship organized at IIT Kanpur. This is a CNN based image recognition model made on keras and tensorflow. The dataset was quite small and the overall validation accuracy obtained was 89% on the validation set.

Agriculture-Commodities-Prices-Seasons

This project is portryal of applied statistical techniques with python and a great deal of time series analysis.

Jupyter Notebook

Clustering-Financial-Articles

Problem description A financial institution news agency has collected 3000 news articles that relates to several matters of financial importance. Before analyzing these unlabeled news, it is only fair to try to partition them into some sort of logical groupings based on their similarities. Your task is to use appropriate unsupervised machine learning algorithm to form the news clusters based on their similarity. Prior to clustering it is recommended to perform basic natural language processing steps such as stemming, tokenization and word vectorization for best results. Notes to keep in mind: There are no duplicate rows in the dataset. Cluster number should start from 0. Data Description There is only one file news.csv that contains date, headlines and text of the news. Column Description id The unique id of the news headline The headline of the news in text text The body of the news in tex

Jupyter Notebook

Self-Driving-Car

Kaggle-Bank-Marketing-Dataset

Mask-No-Mask-Classifier

Jupyter Notebook

Sine-Wave-Outlier-Detection

We have been given 100 samples from a single cycle of a sinusoid. Out of those 100 samples one didnt lie on the sinusoid. Had to detect that. Please go through the notebook, Its very interactive.

Jupyter Notebook

Densenet-on-CIFAR-10

This is the implementation of Densely connected Convolutional Neural Networks trained on CIFAR-10 dataset with standard data augmentation schemes. The result achieved were close to state of the art results.

Jupyter Notebook

Cancer-Dataset

This was a problem statement where we had to determine wheter the type of cancer was benign or malignant. Dataset had few empty values in the column named 'Bare Nuclei', dropped the corresponding rows so as to remove any kind of noise. Two features had a very high value of correlation so dropped one of them. Features were mostly on a similar scale. Finally implemented a Support Vector Classifier with a gaussian kernel and the recall score found on a k fold cross validation tehnique was 1 and the f score was also 1.

Jupyter Notebook

Hackerearth-Criminal-Dataset

Problem Statement - There has been a surge in crimes committed in recent years, making crime a top cause of concern for law enforcement. If we are able to estimate whether someone is going to commit a crime in the future, we can take precautions and be prepared. You are given a dataset containing answers to various questions concerning the professional and private lives of several people. A few of them have been arrested for various small and large crimes in the past. Use the given data to predict if the people in the test data will commit a crime. The train data consists of 45718 rows, while the test data consists of 11430 rows.