Manthan Patel

Here are the Projects, Dashboards and Blogs that I have made/written…

Blogs:

A Blog about How to do installation of TF-1 OD and Object Detection with Pre-Trained TF-1 Model.

A Blog about How to Prepare Annotated Dataset & How to train your own custom model for fruit detection using pre-trained model weights.

Project is a combination of Object detection model and Image classification model, to detect person with no mask and identifying that person to store the data in a downloadable csv file with UI.

For the project, I have done a research on most common harmful and natural ingredients which are used in the Shmapoo.
The Project is a combination of OCR + Text Scraping + Image Scraping + Study regarding common harmful & natural Ingredients in Product

This project is built for classifying total 10 different type of cloths and accessories with VGG Image Classification Model.

This is Kaggle’s annual Machine Learning and Data Science Survey competition for presenting a story of the data science community.

Dataset of X-ray Images are taken from Kaggle
Here, I have manually chosen images of Covid-19 and Normal State.
Entire model is trained on total around 400 images.
For testing, I have used around 50 images.
A Proper explanation of model is also given.
Note: Purpose of this test project is to learn how to build a CNN model. It does not claim any guarantee of anything.

Here, I have used bs4 (BeautifulSoup) library for Extracting Movie/TV Show Data.
MongoDB is used to store the data in database.
Information like ‘Title’,‘Time Period’,‘Rating’,‘Genre’,‘Duration’,‘Votes’,‘Directors’,‘Stars’ and ‘Description’ are extracted.
These information are collected for 60,000+ Movies/TV Shows.
For more understanding of the project, refer IMDB-Movies(60000)-Web-Scraping.ipynb
Finally, I have created Data Visualization with PowerBI for this data.

The Dataset has 2453 rows and 31 columns.
It has 31 columns: ‘having_IP_Address’, ‘URL_Length’, ‘Shortining_Service’, ‘having_At_Symbol’, ‘double_slash_redirecting’, ‘Prefix_Suffix’, ‘having_Sub_Domain’, ‘SSLfinal_State’, ‘Domain_registeration_length’, ‘Favicon’, ‘port’, ‘HTTPS_token’, ‘Request_URL’, ‘URL_of_Anchor’, ‘Links_in_tags’, ‘SFH’, ‘Submitting_to_email’, ‘Abnormal_URL’, ‘Redirect’, ‘on_mouseover’, ‘RightClick’, ‘popUpWidnow’, ‘Iframe’, ‘age_of_domain’, ‘DNSRecord’, ‘web_traffic’, ‘Page_Rank’, ‘Google_Index’, ‘Links_pointing_to_page’, ‘Statistical_report’, ‘Result’
From the Dataset, we have to predict is the website Phishing website or not.
ExtraTreesClassifier has been used for Feature Selection.
I have applied Artificial Neural Network, Random Forest, Decision Tree, K-NN, Naive bayes classification, Logistic Regression and SVM algorithms but at the end, RandomForestClassifier gave better results.

This is a Kaggle Competition Project.
Train and Test Dataset have almost same no. of columns and rows (1460,81).
Various Techniques like Target Guided Encoding, logarithmic transformation technique, StandardScaler. Hyperparameter-tuning etc. have been used.
Algorithms like RandomForestRegressor, Linear Regression, SVR, GradientBoostingRegressor and ANN have been applied.

Dataset has 10000 rows and 8 columns.
We have to predict whether a review is “Positive” or “Negative”.
PortStemmer method has been used for Stemming.
I have also tried WordEmbedding with LSTM.
I have applied many different algorithms LSTM, Bi-Directional LSTM, RandomForestClassifier, MultinomialNB, SVM and KNN.

Message and its final output is separated by Tab space.
From the Dataset, we have to predict the label column:
PortStemmer method has been used for Stemming.
I have applied many different algorithms RandomForestClassifier, MultinomialNB, SVM and KNN.

The Dataset has “PassengerId”, “Survived”, “Pclass”, “Name”, “Sex”, “Age”, “SibSp”, “Parch”, “Ticket”, “Fare”, “Cabin”, “Embarked” columns. It has total around 1300 rows and 12 columns.
From the Dataset, we have to predict the Survived column:
ExtraTreesClassifier has been used for Feature Selection.
I have used Count/Frequency Encoding Technique for Feature Encoding.
I have applied many different algorithms but at the end, KNN gave better results.

The Dataset has ‘density’, ‘pH’, ‘sulphates’, ‘alcohol’, ‘Quality_Category’ columns. It has 4898 rows and 5 columns.
From the Dataset, we have to predict the Quality of Alcohol: “High” or “Low”.
ExtraTreesClassifier has been used for Feature Selection.
I have applied Artificial Neural Network, Random Forest, Decision Tree, K-NN, Naive bayes classification and SVM algorithms but at the end, KNN gave better results.