All Questions

Filter by
Sorted by
Tagged with
0 votes
0 answers
12 views

Extremely poor prediction: Resnet50 Regression

I tried to implement ResNet50 model for Regression prediction. Below is my trial code. This code runs without error. You can also try it without dependency. how to improve so that the model is able to ...
0 votes
0 answers
11 views

Decision tree vs logistic regression feature importances

I have trained Logistic regression and decision tree in skearn on the same standardized dataset (binary classification). Top important coefficients for decision tree are (sorted by tree....
0 votes
0 answers
13 views

General question regarding n_components in UMAP and the resulting plot

Lets say we define n_components in UMAP to be 5 (or really any number greater than 2). When we plot the result in 2D, what features are being displayed? Does UMAP identify the features with the most ...
0 votes
1 answer
9 views

How do I Profile a Jupyter Cell and Rank it by Cumulative Time?

I'm using SciKit-Learn's OneHotEncoder to encode some categorical features. However, the time complexity of the function seems to have an exponential dependence on ...
  • 225
0 votes
1 answer
18 views

Creating function to write excel sheets using pandas

...
0 votes
0 answers
10 views

Best ways to visualize the performance of a regression model?

How best to visualize 10,000 observations of continuous labels against predicted values in a way to summarises the overall performance of the model? I want to avoid using performance metrics such as ...
  • 1
0 votes
1 answer
23 views

how to proportionally round to integer values

I am trying to model student progress. Grades are integer values, but progress predictions are fractional. For example, I have 10 students who are predicted a grade 6, but the model says they should ...
  • 3
0 votes
0 answers
8 views

Do we need padding in global time series models?

In order to forecast product demand for a costumer, I need to predict time series of different lengths. I am using mostly global lightgbm models to do the job. In most use cases, it is common practice,...
0 votes
2 answers
24 views

Classification model- which metric to choose to evaluate

If my project is "Music Genre Classification" Which metric method i need to choose to evaluate, and why ? Thanks a lot :)
  • 11
0 votes
1 answer
9 views

Why do we use multiple recordings from the same set in a dataset?

In the following dataset: https://physionet.org/content/mitdb/1.0.0/ Has been written the following text: "the remaining 25 recordings were selected from the same set to include less common but ...
0 votes
0 answers
8 views

If many features are making a sequential pattern, but individual features are not important, what would be the ideal feature engineering?

I have some extracted features which are sequential. I believe there are some patterns which occurs in sequential manners, for example -- 1,2,3 are inputs, the outputs are 4, but individually they ...
0 votes
0 answers
7 views

Unstable loss in binary classification for time-series data - extremely imbalanced dataset

I am working on deep learning model to detect regions of timesteps with anomalies. This model should classify each timestep as possessing the anomaly or not. My labels are something like this: ...
  • 73
0 votes
1 answer
18 views

Understanding bootstrapping in bias variance decomposition

I was going through bias and variance tradeoff article and it makes use of bias_variance_decomp function from mlxtend library. ...
0 votes
0 answers
15 views

Prior probability shift vs oversampling/undersampling imbalanced datasets

I'm trying to understand what prior probability shift (label drift) in data means. If I understand it correctly then it means that distribution of labels in training dataset differs compared to ...
  • 113
0 votes
0 answers
9 views

What is an acceptable value for mean average precision for a custom object detection model based on a pretained model?

I have custom trained two models from pretrained models available on TensorFlow object detection model zoo (ssd mobilenet and ssd resnet). The objective is to detect objects on a football field. There ...
0 votes
0 answers
5 views

Similar architecture to Siamese

Siamese has been used a lot in different applications starting from text clone detection, pair of similar cloths detection, and many others. It's based on distance metric and a threshold. Are there ...
  • 197
0 votes
0 answers
30 views
+50

CNN for time series: Input 0 of layer "conv2d_5" is incompatible with the layer: expected min_ndim=4, found ndim=2. Full shape received: (None, 2)

I am trying to use CNN on multivariate time series instead the most common usage on images. The number of features are between 90 and 120, depending on which I need to consider and experiment. This is ...
0 votes
0 answers
7 views

What is the difference between freezing & unfrezzing when it come to feature extraction with Transfer Learning

Reading this paper they mention the following In the context of the fine-tuning approach, all of the layers were frozen, except for the last layer, which was removed as it was employed to classify ...
2 votes
1 answer
16 views

NER - What advantage does IO Format have over BIO Format

In this paper, the authors say that they used IO schema instead of BIO in their dataset, which, if I am not wrong, means they just tag the corresponding Entity Type or "O" in case the word ...
0 votes
0 answers
17 views

How to use BayesOpt to tune hyperparameters? [closed]

I am new to Bayesopt. I used the following codes to tune hyperparameters for XGB regression. However, I have the following errors: ----> 1 sess = BTBSession(tunable=params, scorer=scoring_func, ...
0 votes
0 answers
18 views

2 neural netowork trained differently but ending up at the same loss means both the model are identical?

I have 2 neural netowks and I am training them in different way. 1 neural network is generating good predictions and ending up with a certain loss. Another neural network is not doing good predictions....
  • 101
0 votes
0 answers
9 views

Search for fundus data obtained with a peek retina, or a D-EYE, or an iNview

I want to develop a machine learning model for eye disease detection via a smartphone. For this I am looking for data. Especially images of the back of the eye obtained with a Peek retina, a D-eye or ...
0 votes
0 answers
13 views

Imputing values for age of property

I have a dataset with property values for a variety of different types of property. I have the age (from build year) of each property however some of the properties are simply "undeveloped ...
0 votes
1 answer
18 views

Gensim: create a dictionary from a large corpus without loading it in RAM?

The topic modelling library Gensim offers the ability to stream a large document instead of storing it in memory. Streaming is possible for the stage of converting the corpus to BOW, but the ...
  • 23.8k
2 votes
1 answer
47 views

Transformer XL - understanding paper's illustration

If I understand correctly, the Key hidden layer in the Transformer XL is of size 2L * d, where ...
  • 121
0 votes
0 answers
12 views

ANN Input vector representing a network of contests's outcomes between thousands of individuals?

Let's say I have 1000 participants within a tournament, each of the participants having a single ID. One instance of this tournament is a 1 vs 1 contest when one wins and the other loses. Now let's ...
  • 1
2 votes
1 answer
23 views

How to use categorical data in forecasting with Prophet?

I'm trying to create a model to predict the number of players on a video game at a certain time and was trying to figure out how to integrate categorical data into my forecasting problem. So far, my ...
0 votes
0 answers
12 views

How to encode data set with all categorical predictors

I have a data set with all categorical predictors. They are 14 in number. If I do one-hot encoding, I would be getting more than 35 new features, which I think is not right due to the curse of ...
0 votes
1 answer
12 views

How to carry metadata with data examples and labels in python?

Is there a pythonic way to carry around metadata that describes the training examples, such that it preserves (i.e., order) after shuffling and splitting (train/test)? ...
0 votes
0 answers
10 views

Can I fit my model on whole dataset without splitting it to train and test sets to get SHAP values?

I have a small dataset with 297 sample, before using SHAP technique you must have a fitted model, can I fit my model on whole dataset and use it for SHAP technique? I don't want split my dataset to ...
  • 3
0 votes
0 answers
9 views

How to capture a column of a data type

I want to put 2nd column in an array of the following data like [1,1,1,1,..]: ...
  • 1
0 votes
0 answers
6 views

Trying to load csv in pandas, there are 14 columns, but show all in one

céréales = pd.read_csv("../Data/céréales.csv",sep=';') print("la taille de la table céréales est:", céréales.shape) céréales.sample(15) Load data
  • 1
0 votes
0 answers
7 views

Understanding model behavior

I am checking the accuracy of a new model (Kmeans with Iterative RF) over SICE (Single Imputation Of Chained Equation) for missing data imputation. For 5-10% SICE is performing well thereafter our ...
  • 1
0 votes
0 answers
8 views

Machine learning with mixed variables dataset (numerical, categorical and embeddings)

I'm working on a machine learning project where I'm trying to predict the revenue of a movie. My dataset contains mixed data types. There are numerical features (rating, number of votes, release year,....
0 votes
1 answer
11 views

How can I let Non-Transformed Values through Scikit-Learn's Column Transformer?

I'm working with a heterogenous dataset that includes numerical and categorical columns. I want to use Scikit-Learn's ColumnTransformer to ...
  • 225
0 votes
0 answers
14 views

Understanding Cost-sensitive Decision Trees Behaviour

I have a binary classification problem with imbalanced data and am attempting to use cost-sensitive learning to handle the imbalance. I have used LogisticRegression, LinearSVC, SVC and DecisionTree ...
  • 355
2 votes
2 answers
31 views

Difference in proportion between labelled sample and the population

I'm working on a project to predict bots from legit users from social medias. The data that I collected has about 5% of bots for 95% of legit users. The problem is as I labelled my data, I was more ...
  • 212
0 votes
0 answers
12 views

Stacked ensemble model

In regression, multicollinearity between variables would need to be removed to suit the model assumption. In building a stacked ensemble model, with say SVM, xgb and a decision tree as a base model ...
0 votes
0 answers
9 views

not enough values to unpack (expected 4, got 1)

I am trying to transform the data values from .csv file to python using sklearn. The problem that I encountered is that it generates this error: Exception has occurred: ValueError not enough values ...
0 votes
1 answer
35 views

How to train AI to recognize a Snapchat's bitmoji's gender

I would like to enable my Python script to get a vague idea of a bitmoji's gender. I found these two git repositories that are able to differentiate between them. How would I use them to have my ...
0 votes
0 answers
34 views

Summerize features for time-series predicted result

Background Our project is to classify if the given meter is EV or not, we got time-series data for each meter, got columns like time(30min intervals), and consumption in power. We also have some users ...
-1 votes
0 answers
9 views

ModuleNotFoundError: No module named 'm0c_logistic_regression'

I want to use import this model for prediction, but It returns an error message on jupyter notebook. ...
0 votes
0 answers
25 views

Kolmogorov-Smirnov hypothesis test

I am using this package to test the similarity of two distributions. It says: The one-sample test compares the underlying distribution F(x) of a sample against a given distribution G(x). The two-...
0 votes
1 answer
18 views

Why does Adam outperform SGD in logistic regression?

I am training a logistic regression model. In case it matters, the features are 1376-dimensional embeddings output from a neural network. I tried both SGD and Adam with a learning rate of $10^{-3}$ ...
  • 113
0 votes
0 answers
48 views

Difference between CNN filter and channel pruning

I've been reading up on DNN pruning techniques and came across filter pruning and channel pruning in CNNs. However, as was my understanding, confirmed by this paper: Pruning a filter in layer $i$ is ...
  • 101
0 votes
1 answer
18 views

How to choose neural network architecture for a relatively small dataset with less than 10 features for regression?

How to go about selecting an architecture for a dataset with 80 datapoints and 9 features for a regression model? Working on the Desarhnais dataset, with "Effort" as the target variable. ...
0 votes
0 answers
18 views

Number of CNN's for processing timeseries using 2D CNN

I am processing time-series for classification using 2D CNN model (single channel) where I have converted stationary time series data into 2D image using an "imaging algorithm" known as ...
0 votes
1 answer
20 views

Does Scikit-Learn's OneHotEncoder make all Columns Categorical?

I've been using Scikit-Learn's OneHotEncoder to turn categorical data into binary columns, however, it seems that fitting ...
  • 225
0 votes
0 answers
18 views

which code is right to do the times series analysis in this data set?

The dataset has two variables that are date (Jan 2018 to Sep 2019) and the value ($million). It concerns the value amount of air import and I would like to do time series analysis. I am a university ...
0 votes
0 answers
8 views

Getting matrix multiplication error when running inference on a Pytorch model(LeNet) using MNIST data

I have successfully trained the Pytorch LeNet model on the MNIST data, but have a probelm when running inference on single images. Below is my model architecture: ...

15 30 50 per page