Making analysis for backpack with different brands and features, and making model for price prediction using Backpack Data. Model uses combinations of several models, CATboostRegressor, LGBMRegressor, XGBRegressor with their optimal parameters from OPTUNA and linear regressor as meta learner.

Sticker sale prediction using Forecasting Sticker Sales Data and LGBMRegressor with optimized parameters from OPTUNA.

Making content based recomendation system with combinations of cosine similary and Euclidean distance formula by using game data Metacritic Games 1995-2024. Data collected from metacritic with code Metacritic Data Collect.

Exploratory data analysis and data cleaning for Metacritic Games 1995-2024 data.

BS4 code for collecting data from metacritic. Code collects all games for all platforms, but it can be easily modified for specific searches.

Exploratory data analysis, data cleaning and basic collaborative filtering restaurant recomendation system for NYC Restaurants Data - Food Ordering and Delivery data.

Filtering players according to their roles, picking and grouping players in pick range. Comparing groups to find one with best composition.

Classification of images of American Sign Language with CNN model, American Sign Language Dataset and augmented forms of images are used for both training and testing.

Exploratory data analysis, NLP and sentiment analysis by using multi-channel CNN for Sentiment Analysis for Mental Health Data.

Sentiment analysis with comments collected from pre selected TWS products at Hepsiburada.com.

Code for Turkish text summarization. Code generate extraction of given text via using TextRank alagorithm and FastText vord data.

Code that allows you to use Google Finance via python, also allows you to use both Yfinance and GoogleFinance. For some reasons, yfinance returns empty data one one day data, code tries yfinance, in the absence of data uses Google Finance.

Collecting data for symbols being IPO in 2 years, making analysis and creating model using statsmodel and backward elimination to find initial count of hype days.

Collecting data for symbols which are active in given sector area, finding optimal pick of group of symbols with optimal weights that aims high gain and low risk.

Collecting data for symbols, and after some filtering calculations predicting price of filtered stocks with using MonteCarlo and LSTM.

The methods of supergravity allow us to derive a multi-field F-term potential. Using this, we denote a generic and non-positive single-field F-term potential. We insert our theory into the scalar-gravity part of the (2,1|1) invariant superconformal action. That action leads us to a de Sitter solution at the inflationary trajectory. One can denote stabilization of fields in terms of the Kähler kinetic terms and single-field slow-roll inflation parameters. We combine these with the de Sitter swampland conjecture to generate a bounded conjecture. This approach allowed us to show that the single field slow-roll inflation works in harmony with bounded de Sitter conjecture for any concave inflation potential.

We provide a tensor calculus for n-number of N=(1, 0)linear multiplets in six dimensions. The coupling of linear multiplets is encoded in a function F_ij that is subject to certain constraints. We provide various rigid and local supersymmetric models depending on the choice of the function F_ij and provide an interesting off-diagonal superinvariant, which leads to an R^2 supergravity upon elimination of auxiliary fields.