11
votes
Accepted
Learning with Positive labels only
The topic you are interest in is called "PU learning" or "positive and unlabeled learning".
You can start by having a look into survey literature.
- 1,024
10
votes
Accepted
8
votes
Accepted
Sklearn: unsupervised knn vs k-means
Unsupervised k-NN
Unlike k-means, the unsupervised k-nn does not associate a label to instances. All it can do is tell you what instances in your training data is k-nearest to the point you are ...
- 8,876
8
votes
Accepted
Coordinate System's influence on $L$ distances (Manhattan and Euclidean)
For example, consider the green line. What is its length? In $L_2$, the answer is $1$, in $L_1$, the answer is $1$ as well.
Now, for the same line, let's rotate it $45^\circ$ counterclockwise. What ...
- 3,003
7
votes
What is difference between Nearest Neighbor and KNN?
Not really sure about it, but KNN means K-Nearest Neighbors to me, so both are the same. The K just corresponds to the number of nearest neighbours you take into account when classifying.
Maybe what ...
- 626
6
votes
Accepted
Is it sensible to use the ROC curve with an KNN model? And if so why?
ROC curves (and the AUC metric) are used for evaluating model performance on a classification task. If you use KNN for classifying, then you can evaluate your model on it.
Probability, in the context ...
- 451
5
votes
Accepted
How Does Weighted KNN Work?
We can view nearest neighbor as a voting process where we consult our $k$ nearest neighbor.
We give the $i$-th data point a voting weight $w_i$.
In your example, each data point in class $A$ has ...
- 3,003
5
votes
Is Annoy a machine learning algorithm to find nearest neighbor ? and is it similar to K nearest neighbor algorithm?
Don't have enough reputation to comment to a resource, so answering this myself.
About Annoy
Annoy is a library being used here for finding approximate nearest neighbours, approximate being the key ...
- 151
5
votes
When using KNeighborsClassifier, what is the motivation of using weights="distance"?
weights = 'distance' is in contrast to the default which is weights = 'uniform'. When weights are uniform, a simple majority ...
- 21.2k
4
votes
Accepted
How does sklearn KNeighborsClassifier compute class probabilites?
The class probabilities are the normalized weighted average of indicators for the k-nearest classes, weighted by the inverse distance.
For example: Say we have 6 classes, and the 5 nearest examples ...
- 2,381
4
votes
Accepted
How does KNN work if there are duplicates?
Your reasoning is correct - you should consider duplicate points as separate. You can see that this must be the case in several ways:
Introduction of small random noise to the data should not affect ...
- 2,121
4
votes
Accepted
Is there any way of ordering/sorting vectors?
If I understand your question correctly, you're looking for something that another (beyond KD-trees) standard space partitioning algorithm does. This one is called BSP-trees (for Binary Space ...
- 744
4
votes
Accepted
Is K-NN applicable for binary variables?
Yes, you just have to find a suitable distance metric, instead of using the default Euclidean distance. Euclidean distance will work, but it loses a lot of its positive points when used on a non-...
- 937
4
votes
Accepted
Evaluation metrics for Decision Tree regressor and KNN regressor
Generally when ever we are trying to compare between models and to choose the best one, we go for other metrics like AIC, BIC, AUC(this is not applicable as it is used for classification algorithm) ...
- 2,392
4
votes
Accepted
Why does the overfitting decreases if we choose K to be large in K-nearest neighbors?
Overfitting is "The production of an analysis which corresponds too closely or exactly to a particular set of data, and may therefore fail to fit additional data or predict future observations ...
- 335
4
votes
What is difference between Nearest Neighbor and KNN?
Scikit wrote in his documantation:
sklearn.neighbors provides functionality for unsupervised and supervised neighbors-based learning methods. Unsupervised nearest neighbors is the foundation of many ...
3
votes
How to calculate the silhouette coefficient?
a(i) : the average distance between 'i' and all other data within the same cluster (source)
b(i) : the lowest average distance of 'i' to all points in any other cluster, of which 'i' is not a member ...
- 1,282
3
votes
Accepted
Knn and euclidean distance
You can think of examples as vectors in $\mathbb{R}^p$, where $p$ is the number of features. Two examples will be very similar if the distance between them is close to $0$ (in the extreme case, if two ...
- 6,051
3
votes
Accepted
sklearn.neighbors.NearestNeighbors - knn for unsupervised learning?
The unsupervised version simply implements different algorithms to find the nearest neighbor(s) for each sample.
The kNN algorithm consists of two steps:
Compute and store the k nearest neighbors ...
- 6,347
3
votes
Accepted
Why do I not get 100% Accuracy with KNN with $K=1$
In your training set (X_2,y), there are some samples with the same input features X_2 but different labels y. For example, the 73rd and 147th samples, which are labelled into class 0 and 1, ...
- 185
3
votes
K-Nearest Neighbours algorithm explanation needed
K Nearest Neighbors is a Classification Algorithm. Just as with every classification algorithm is important that the algorithm doesn't "remember" the answers and that the answer it gets can be ...
- 2,302
3
votes
How to save a knn model?
Importing the library
from sklearn.externals import joblib
Saving your model after fitting the parameters
...
- 2,074
3
votes
Is the distance in Nearest Neighbors a good measure of similarity?
This is counter-intuitive, because one would expect [0,2,0,0] to be more similar to [0,1,0,0] than [0,1,1,0].
No this is expected, since the two points are exactly at the same distance in the ...
- 25.3k
3
votes
Accepted
Calculating distance between data points when there are more than 3 features in KNN algorithm
No, you can definitely search for k-NN with more than 2-dimension data. Here is an example based on sklearn:
...
- 1,395
3
votes
Accepted
Optimal selection of k in K-NN
Your idea isn't wrong, however in k-NN there always might be a case where you have the same number of votes for 2 or more classes (e.g. you have $k=6$ and you have 3 samples of one class vs 3 of ...
- 7,978
3
votes
Optimal selection of k in K-NN
As I was playing with this problem, Djib just wrote an answer which is certainly better than whatever I could have come up with. To illustrate Djib's point, here is a small demonstration that as soon ...
- 25.3k
2
votes
Why is cross-validation score so low?
I know this question has been here for two years, however, I was having the same problem when using cross_val_score on my data and I ended up here.
The results ...
- 21
2
votes
Accepted
Why is cross-validation score so low?
In your random forest, this is due to the fact that your final model is overfitting. Sklearn's GridSearchCV has a default argument ...
- 6,051
2
votes
sklearn.neighbors.NearestNeighbors - knn for unsupervised learning?
The confusion comes from the way Sklearn designed their code.
Short answer
The "unsupervised" version you mention is not a K-Nearest Neighbour algorithm (which is implemented here). In its ...
- 6,026
2
votes
Accepted
scalable tools to build kNN graph over sparse data
I think what you might be looking for is,
L2Knng: Fast Exact K-Nearest Neighbor Graph Construction with L2-Norm Pruning
They have multiple runtime options specifically for different kinds of ...
- 111
Only top scored, non community-wiki answers of a minimum length are eligible