#pragma GCC target("avx2")
#pragma GCC optimize("O3")
#pragma GCC optimize("unroll-loops")

#include <iostream>
#include <vector>
#include <cassert>
#include <atcoder/modint>
#include <atcoder/convolution>

using namespace std;
using namespace atcoder;
using mint = modint998244353;
// verify: https://judge.yosupo.jp/submission/153840
vector<mint> fps_inv(vector<mint> f,int deg = -1){
    if(deg==-1) deg = f.size() - 1;
    assert(f.size());
    assert(f[0].val()!=0);
    vector<mint> g = {(mint)1/f[0]};
    for(int i=1;i<=deg;i<<=1){
        vector<mint> h = convolution(g,f);
        h.resize(2*i);
        for(mint &u:h) u = -u;
        h[0] += 2;
        g = convolution(g,h);
        g.resize(2*i);
    }
    g.resize(deg + 1);
    return g;
}

vector<mint> fps_derivative(vector<mint> &f){
    vector<mint> ret;
    for(int i=1;i<f.size();i++){
        ret.push_back(f[i]*i);
    }
    if(f.size()==1) ret.push_back(0);
    return ret;
}

// get integral f(x) dx, s.t. const = 0;
vector<mint> fps_integral(vector<mint> &f){
    vector<mint> g(f.size());
    for(int i=g.size() - 1;i>=1;i--){
        g[i] = f[i - 1]/(mint)(i);
    }
    g[0] = 0;
    return g;
}

// get log(f) = -sum (1 - f)^n/n 
vector<mint> fps_log(vector<mint> f,int deg = -1){
    if(deg==-1) deg = f.size() - 1;
    assert(f[0].val()==1);
    vector<mint> g = convolution(fps_derivative(f),fps_inv(f,deg));
    g.resize(deg + 1);
    return fps_integral(g);
}

// get exp(f) = sum f^i/i!
vector<mint> fps_exp(vector<mint> f,int deg = -1){
    if(deg==-1) deg = f.size() - 1;
    assert(f.size());
    assert(f[0].val()==0);
    vector<mint> g = {(mint)1};
    for(int i=1;i<=deg;i<<=1){
        vector<mint> h(2*i);
        h[0] = 1;
        vector<mint> log_g = fps_log(g,h.size() - 1);
        for(int j=0;j<h.size();j++) h[j] -= log_g[j];
        for(int j=0;j<min(f.size(),h.size());j++) h[j] += f[j];
        g = convolution(g,h);
        g.resize(2*i);
    }
    g.resize(deg + 1);
    return g;
}

using ll = long long;
const int MX = 1000010;
mint f[MX],inv[MX],fi[MX];
constexpr ll mod = 998244353;
void solve(){
    inv[1] = 1;
    for(int i=2;i<MX;i++){
        inv[i] = mod - (mod/i)*inv[mod%i];
    }
    f[0] = fi[0] = 1;
    for(int i=1;i<MX;i++){
        f[i] = f[i-1]*i;
        fi[i] = fi[i-1]*inv[i];
    }
}

// input f(x) output g(x) = f(x + d)
// 階乗の逆元の計算が必須
vector<mint> shift(vector<mint> ff,int d){
    int i,len = ff.size();
    vector<mint> f1(len + 1),f2(len + 1);
    mint x = 1;
    for(i=0;i<len;i++){
        f1[i] = ff[i]*f[i];
        f2[len - i] = x*fi[i];
        x *= d;
    }
    vector<mint> f3 = convolution(f1,f2);
    vector<mint> ret;
    for(i=len;i<2*len;i++){
        ret.push_back(f3[i]*fi[i - len]);
    }
    return ret;
}

// input: s[0] + s[1]x + ... s[n - 1]x^n - 1
// output: (c[0] - c[1] - .. c[d])s = 0かつc[0] = 1なるcを存在すればどれか一つ
// copied from: https://nyaannyaan.github.io/library/fps/berlekamp-massey.hpp
// verified: https://judge.yosupo.jp/submission/199848
// algorithmの気持ち
// deg(c) = kの時に s[i] + Σ_{j<k} c[j]*s[i - 1 - j]　を計算する
// 0ならOKで、0じゃないときは、1個前のcの線形結合で補正項を入れる
// 今までは成功していてたので、1個前のcの線形結合で補正したら、自分より下はつじつまが合うし、補正を入れるので徐々にあっていく
template <typename mint>
vector<mint> BerlekampMassey(const vector<mint> &s) {
    const int N = (int)s.size();
    vector<mint> b, c;
    b.reserve(N + 1);
    c.reserve(N + 1);
    b.push_back(mint(1));
    c.push_back(mint(1));
    mint y = mint(1);
    for(int ed = 1; ed <= N; ed++){
        int l = int(c.size()), m = int(b.size());
        mint x = 0;
        for (int i = 0; i < l; i++) x += c[i] * s[ed - l + i];
        b.emplace_back(mint(0));
        m++;
        if(x == mint(0)) continue;
        mint freq = x / y;
        if(l < m){
            auto tmp = c;
            c.insert(begin(c), m - l, mint(0));
            for (int i = 0; i < m; i++) c[m - 1 - i] -= freq * b[m - 1 - i];
            b = tmp;
            y = x;
        }else{
            for (int i = 0; i < m; i++) c[l - 1 - i] -= freq * b[m - 1 - i];
        }
    }
    reverse(begin(c), end(c));
    return c;
}

// BostanMori
// input: P(x)/Q(x),deg(P(x))<deg(Q(x)),
// output: [x^N](P(x)/Q(x)) 
// O(dlog(d)logN), (d := deg(Q(x)))
// TODO: MSB-firstにするともっと早いらしい
// verified: https://judge.yosupo.jp/submission/199853
mint BostanMori(vector<mint> P,vector<mint> Q,ll N){
    while(N){
        vector<mint> _Q(Q.size());
        for(int i=0;i<Q.size();i++) _Q[i] = i&1 ? -Q[i] : Q[i];
        vector<mint> nP = convolution(P,_Q);
        vector<mint> nQ = convolution(Q,_Q);
        Q.resize(nQ.size()/2 + 1);
        for(int i=0;i<nQ.size();i+=2) Q[i/2] = nQ[i];
        if(N&1){
            P.resize(nP.size()/2);
            for(int i=1;i<nP.size();i+=2) P[i/2] = nP[i];
        }else{
            P.resize((nP.size() + 1)/2);
            for(int i=0;i<nP.size();i+=2) P[i/2] = nP[i];
        }
        N /= 2;
    }
    return P[0]/Q[0];
}

mint pw(mint a,ll x){
    mint ret = 1;
    while(x){
        if(x&1) (ret *= a);
        (a *= a); x /= 2;
    }
    return ret;
}

// f^k mod x^N をsparceなfに対して計算
// [x^0]f \neq 0 を仮定
// fの非零項がm個で、O(mN) になる
// verified: https://atcoder.jp/contests/nadafes2022_day1/submissions/52160796
vector<mint> sparse_pow(vector<mint> &f,int k,int N){
    assert(f.size());
    assert(f[0]!=0);
    mint iv = pw(f[0],mod - 2);
    vector<pair<int,mint>> ff;
    for(int i=1;i<f.size();i++){
        if(f[i].val()) ff.push_back({i,f[i]});
    }
    vector<mint> ret(N);
    ret[0] = pw(f[0],k);
    for(int i=1;i<N;i++){
        for(int j=0;j<ff.size();j++){
            int deg = ff[j].first;
            if(deg>i) break;
            ret[i] += ff[j].second*deg*ret[i - deg];
        }
        ret[i] *= k;
        for(int j=0;j<ff.size();j++){
            int deg = ff[j].first;
            if(deg>i) break;
            ret[i] -= ff[j].second*(i - deg)*ret[i - deg];
        }
        ret[i] *= inv[i];
    }
    return ret;
}

int main(){
    int p1,p2,q1,q2,t; cin >> p1 >> p2 >> q1 >> q2 >> t;   
    mint p = (mint)p1/(mint)p2; mint q = (mint)q1/(mint)q2;
    vector<mint> g(t + 2),h(t + 1);
    mint pr = 1,a = q;
    for(int i=0;i<=t;i++){
        g[i + 1] = -pr;
        h[i] = pr;
        pr *= a; a *= q;
    }
    g[0] = 1;
    for(int i=1;i<=t;i++) g[i] *= p;
    vector<mint> f = convolution(fps_inv(g),h);
    cout << f[t].val() << "\n";
    // for(int i=0;i<=2;i++) cout << (g[i]*8).val() << " ";
    // cout << "\n";
    // for(int i=0;i<=2;i++) cout << (h[i]*64).val() << " ";
    // cout << "\n";
}